Salazar QG

UNIVERSIDAD NACIONAL DE INGENIERÍA
FACULTAD DE CIENCIAS
TESIS
“DEVELOPMENT IN C++ AND PYTHON OF A TIME
STRUCTURE ANALYSIS TOOL FOR PARTICLE BEAM
ANALYSIS FOR THE MINERvA TEST BEAM EXPERIMENT”
PARA OBTENER EL GRADO ACADÉMICO DE MAESTRO EN

CIENCIAS CON MENCION EN FÍSICA
ELABORADO POR:
GERALD FERNANDO SALAZAR QUIROZ
ASESOR:
Dr. CARLOS JAVIER SOLANO SALINAS
LIMA – PERÚ
2016
i
Dedicado a Fisher, Arnold, Ana, Nataly, Abel y Carmen.

ii
Do not go gentle into that good night,

Old age should burn and rave at close of day;
Rage, rage against the dying of the light.
Though wise men at their end know dark is right,

Because their words had forked no lightning they
Do not go gentle into that good night.
Dylan Thomas, 1914 - 1953

iii
Acknowledgments
How can anyone resume all the time, help, advice and support received over
the time that this thesis was done in a couple of lines?. For all of them thanks,
even the tiny hint was useful and well taken. However, there are some important
persons and institutions that I do want to thanks.
To my advisor Dr. Javier Solano for his support and advice during all the time
I spent at Fermilab, as well to Dr. Jorge Morfin: great listener and even greater
physicist. To all the MINERvA Collaboration and specially to Deborah for their
kindest support. From the very first moment in the experiment I was able to see
by myself that science is truly a hard, colaborative and open endeavour in the
pursue of knowledge.
To the Test Beam crew: Leo Bellatoni for all the guidance, Aaron Bercellie,
Rob Fine, Anne Norrick and Geoff Savage, as well to Howard Budd for the count-
less hours teaching me the basis of how to be a Detector Expert. To Edgar
Valencia, from who I receive my first hints in the painful walk of learning ROOT,
which by the way it is not as hard as once appered specially when my code was
crashing even that, minutes before it was working okey. Of course I know the
answer to that: bad use of pointers.
To Paola from who I am still learning a lot of great things. To my 5 Sauk
Circle’s friends: Aaron, Anu, Anna and Maya. During the time I was writing this
thesis my mind was driven gently to the awesome time we spent together, and
finally to Mayra (#FA’s for the win).
iv
Abstract
We present in this thesis the results of the design and programming of a temporal anal-
ysis tool for experiment data MINERVA Test Beam (TbTaTool), which receives pions and
electrons in the energy range corresponding to the end of the interactions of neutrinos
energy states with MINERvA detector. The TbTaTool is independent, flexible and adapt-
able to other contexts where we need to analyze the distribution of events over time,
because it has separated datasets’ production from the tool itself. Our tool has been
applied to the data (Run 2 and Run 3) that the experiment has obtained in MTest at Fer-
milab, focusing on the variables of frequency of spill of the particles (MI’s spill frequency),
duration of spill (MI’s spill duration) and the profile over time of packets corresponding to
the delivery of the particles (time profile). Our calculations show 0.01% of difference for
the frequency of spill of the particles and 9.34%, for the second variable, compared to the
values indicated by the Fermilab’s Accelerators Division. Furthermore, this tool had set
the basis for constructing a real time DAQ visualizator of the measurement process.
Resumen
Se presenta los resultados del diseño y programación de una herramienta de análisis
temporal para los datos del experimento MINERvA Test Beam (TbTaTool), que recibe
piones y electrones en el rango de energı́a correspondiente a los estados energéticos
finales de las interacciones de neutrinos con el detector MINERvA. El TbTaTool es inde-
pendiente, flexible y adaptable a otros contextos donde se quiera analizar la distribución
de eventos respecto al tiempo. Nuestra herramienta ha sido aplicada a los datos (Run 2 y
Run 3) que el experimento ha obtenido en el MTest en el Fermilab, enfocándonos en las
variables de frequencia de entrega de las partı́culas (MI’s spill frequency), duración de
la entrega (MI’s spill duration) y el perfil en el tiempo de los paquetes correspondientes
a la entrega de las partı́culas (Time Profile). Nuestros cálculos muestran un 0.01% de
diferencia para la frecuencia de entrega de las partı́culas y 9.34%, para la segunda vari-
able, frente a los valores indicados por la División de Aceleradores de Fermilab. Además
esta herramienta ha establecido el fundamento para la construcción de visualizador del
proceso de adquisicion de datos en tiempo real.
Contents
1 Theoretical Framework 2
1.1 Weak Interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 Neutrino Physics . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.2.1 Solar and Atmospheric Neutrinos . . . . . . . . . . . . . . . 8
1.2.2 Neutrino oscillations . . . . . . . . . . . . . . . . . . . . . . 11
2 Neutrino Experiments 14
2.1 Neutrino interaction with matter . . . . . . . . . . . . . . . . . . . 15
2.1.1 Intermediate Energy Cross Sections . . . . . . . . . . . . . 17
2.2 Neutrino Cross Sections Measurements . . . . . . . . . . . . . . . 22
2.3 Neutrino experiments . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.3.1 T2K experiment . . . . . . . . . . . . . . . . . . . . . . . . 26
2.3.2 MiniBoone . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
3 MINERvA a cross-section neutrino experiment 29

3.1 ”Bringing Neutrinos into Sharp Focus” . . . . . . . . . . . . . . . . 30
3.1.1 The NuMI Beam at Fermilab . . . . . . . . . . . . . . . . . 31
3.1.2 MINERvA Detector . . . . . . . . . . . . . . . . . . . . . . . 32
3.2 Test Beam Detector . . . . . . . . . . . . . . . . . . . . . . . . . . 38
3.2.1 Why is necessary to have a Test Beam program . . . . . . 40
3.2.2 FTBF Beam production . . . . . . . . . . . . . . . . . . . . 41
3.2.3 Test Beam detector’s auxiliary systems . . . . . . . . . . . 42
3.2.4 How an event is recorded: trigger logic and DAQ for MIN-
ERvA Test Beam 2 . . . . . . . . . . . . . . . . . . . . . . . 50
v
CONTENTS vi
4 Time structure of the FBTF Beam 54

4.1 General Concepts in Accelerator Physics . . . . . . . . . . . . . . 55
4.1.1 Fermilab’s accelerator complex . . . . . . . . . . . . . . . . 57
4.2 Radio Frequency Systems (RF Systems) . . . . . . . . . . . . . . 59
4.2.1 RF Cavities . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
4.2.2 Power Losses in Cavity . . . . . . . . . . . . . . . . . . . . 63
4.3 Time structure of the Beam . . . . . . . . . . . . . . . . . . . . . . 64
4.3.1 Synchronicity condition . . . . . . . . . . . . . . . . . . . . 64
4.3.2 synchrotron Oscillation, Buckets and Bunchs . . . . . . . . 66
4.3.3 Time Structure of the Beam at MCenter . . . . . . . . . . . 67
4.3.4 The Booster and the formation of Buckets and Batchs . . . 68
4.3.5 Main Injector: formation of MI Cycle, Resonance Extraction
and Spill frequency . . . . . . . . . . . . . . . . . . . . . . . 69
5 Tools for Data Analysis 71

5.1 General concepts of Statistical Analysis . . . . . . . . . . . . . . . 72
5.1.1 ROOT: Data Analysis Framework . . . . . . . . . . . . . . . 73
5.2 Description of the Spill scale in the beam at MTest . . . . . . . . . 78
5.2.1 Description of the data . . . . . . . . . . . . . . . . . . . . . 79
5.3 Development and Features of the Time Analysis Tool (TbTaTool) . 80
5.3.1 Election of a reference point in time . . . . . . . . . . . . . 82
5.3.2 Number of spills . . . . . . . . . . . . . . . . . . . . . . . . 84
5.3.3 Implementation of the tool for the Test Beam data . . . . . 86
6 Results 89
6.1 Time profile of the MTest Beam . . . . . . . . . . . . . . . . . . . . 89
6.2 Main Injector’s Spill Frequency for Run2 and Run3 . . . . . . . . . 93
6.3 Main Injector’s Spill duration at MTest for Pions and Electrons . . . 94
7 Conclusions 96
CONTENTS vii
A Auxiliary plots 100

A.1 Spill frequency for only Pions (Run 2) . . . . . . . . . . . . . . . . 100
A.2 Spill frequency for Electrons (Run 3) . . . . . . . . . . . . . . . . . 103
A.3 Spill duration for only Pions (Run 2) . . . . . . . . . . . . . . . . . 105
A.4 Spill duration for only Electrons (Run 3) . . . . . . . . . . . . . . . 107
B Participation in the MINERνA experiment during 2015 109

B.1 Shifts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
B.2 Detector Expert Training . . . . . . . . . . . . . . . . . . . . . . . . 111
B.3 PMT Cross Talk problem . . . . . . . . . . . . . . . . . . . . . . . . 112
B.4 PMTs Testings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
B.5 Data and MonteCarlo Rock Muon eye scanning . . . . . . . . . . . 115
B.6 Participation in published articles during 2015 . . . . . . . . . . . . 117
C Radiofrequency Cavities theory 119

C.1 Electromagnetic Waves . . . . . . . . . . . . . . . . . . . . . . . . 119
C.1.1 Electromagnetic waves in Vacuum . . . . . . . . . . . . . . 119
C.1.2 Electromagnetic waves in Matter . . . . . . . . . . . . . . . 121
C.1.3 Electromagnetic Waves in Conductors . . . . . . . . . . . . 122
C.2 Guide Waves and Resonant Cavities . . . . . . . . . . . . . . . . . 123
C.2.1 Resonant Cavities . . . . . . . . . . . . . . . . . . . . . . . 126
C.2.2 Power Losses in Cavity: Q of Cavity . . . . . . . . . . . . . 129
C.3 ToF basic physics . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
D Documentation of Code 132

D.1 Subroutines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
D.1.1 Numbers of events in one subrun . . . . . . . . . . . . . . 132
D.1.2 Numbers of events for a Run . . . . . . . . . . . . . . . . . 132
D.1.3 Extracting the timestamp of one subrun . . . . . . . . . . . 133
D.1.4 Extracting the timestamp for the first event of all spills inside
one subrun . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
D.1.5 Calculation of the Kick-Off of One Subrun . . . . . . . . . . 134
CONTENTS viii
D.1.6 Generation of txt file with timestamps of beginning of the

subrun . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
D.1.7 Matching the two closest timestamps . . . . . . . . . . . . . 134
D.1.8 Matching two points for a set of subruns during a data run . 135
D.1.9 Generate a matched times analysis . . . . . . . . . . . . . 135
D.1.10 Plotting one variable against all energies of the data set . . 136
D.1.11 Plotting stacked histograms for all energies . . . . . . . . . 136
D.1.12 Plotting one variable for polarities . . . . . . . . . . . . . . . 136
D.1.13 Generating histograms for one variable for all energies and
stacked them in one plot . . . . . . . . . . . . . . . . . . . . 137
D.1.14 GetValuesHistograms . . . . . . . . . . . . . . . . . . . . . 137
D.1.15 Production of all the analysis . . . . . . . . . . . . . . . . . 138
E Developed Code 139

E.1 TbTaTool Time Profile (for only one subrun) . . . . . . . . . . . . . 139
E.2 TbTaTool Spill Frequency and Spill Duration (for only one subrun) . 148
E.2.1 Cut a variable of time into different parts according to a criteria149
E.2.2 Calculation of different variables regarding time . . . . . . . 149
E.3 Auxiliary Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
E.3.1 Tool for match the $39 signal and the corresponding root file 150
E.3.2 Conversion between Unixtime into readable human time . . 151
E.4 Election of the reference point for time profile . . . . . . . . . . . . 153
E.5 Creation of ROOT Files from TXT files . . . . . . . . . . . . . . . . 154
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
List of Figures
1.1 The 2015 Nobel Prize in Physics winners: Takaaki Kajita (left) and
Arthur B. McDonald (right) . . . . . . . . . . . . . . . . . . . . . . . 4
1.2 Fundamental vertex for weak interactions. . . . . . . . . . . . . . . 5
8figure.caption.7
1.4 (a) Path of the atmospheric neutrinos during its travels through the
earth. (b) Zenith angle events distribution of e-like and µ-like events
in Super-Kamiokande in the range of energy below 1.33GeV[11].
The red line, indicates the best fit for the data points, while the
boxes show the Monte Carlo events expectation considering no
oscillations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.1 Total cross-sections for Charge Current Quasi-elastic Scattering

on Carbon measured by different experiments. This figure has
been taken from the Conference MINERνA 101 by C. Patrick [27] . 16
2.2 Fundamental vertex. This figure has been taken from the Confer-
ence MINERνA 101 by C. Patrick [27] . . . . . . . . . . . . . . . . 16
2.3 Total neutrino and antineutrino per nucleon CC cross-sections di-
vided by neutrino energy and plotted as a function of energy. Data
include the low energy data from N ([6]), ∗ ([]), ([]), and ? ([]).
The three main regions are ploted as quasi-elastic (dashed), res-
onance production (dot-dash), and deep inelastic scattering (dot-
ted). The predicitions for each region are provided by the NUANCE
generator ([7]) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
ix
LIST OF FIGURES x
2.4 All νµ quasi-elastic scattering cross-sections, νµ n → µ− p measure-

ments until the year 2002 by [7] as a function of neutrino energy
for different nuclear targets. . . . . . . . . . . . . . . . . . . . . . . 20
2.5 Data measurements of νµ CC resonant single-pion production re-
ported by experiments with no additional corrections derived of nu-
clear targets or invariant mass selections. The continuous curve
has been generated by NUANCE. [26]. . . . . . . . . . . . . . . . 22
2.6 Absolute coherent pion production measurements from a variety
of nuclear targets and samples for NC and CC data [7, p. 33]. . . 23
2.7 Charge Current Neutrino interactions in the intermediate region
and the experiments that cover those regions. From M. Martin
presented in the NuFact15 [22]. . . . . . . . . . . . . . . . . . . . . 24
2.8 Geographical map of the T2K detection the east coast of Japan.
Images from the ofical page of T2K experiment http://t2k-experiment.
org/. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
2.9 MiniBooNE and MicroBooNE experiments . . . . . . . . . . . . . . 28
3.1 Front view of the MINERvA detector at NuMI Hall at Fermilab. . . . 30

3.2 Diagram of NuMI and its different elements that produce the neu-
trino beam used by MINERvA, MINOS and NOvA experiments. fig.
by . Pavlovic. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.3 Side view of the complete detector showing the nuclear target, the
active region and the surrounding calorimeter regions. Image[2] . 33
3.4 Active Scintillator Modules, and the 5 types of nuclear modules.
Image from A. Norrick [25] . . . . . . . . . . . . . . . . . . . . . . 34
3.5 Images form [2] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
3.6 Images by C.L. McGivern . . . . . . . . . . . . . . . . . . . . . . . 37
3.7 Typical time slice of . Screenshoot taken on 3/9/16 . . . . . . . . . 37
3.8 Images of the FTBT. . . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.9 Calorimetric response for positive (left) and negative (right) pions
for low energies in the Test Beam Program 1 for low energies. . . 40
3.11 Energies and Polarities taken in the Run 1. Tables made by R. Fine. 43
LIST OF FIGURES xi
3.12 Energies and Polarities taken in the Run 2 and 3. Tables made by
R. Fine. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
3.13 The two configuration for the calorimetric regions in the TB2. Im-
age made by A. Bercelli. . . . . . . . . . . . . . . . . . . . . . . . . 44
3.14 (Left) Schematic view of the Test Beam Detector and the Auxiliary
detection System. (Right) Components of the Test Beam and its
functions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
3.15 (a) TDC-CAMAC Lecroy 3377 used in NuTeV and CDF. . . . . . . 47
3.16 Veto system and its six paddles. Photo by A. Norrick . . . . . . . . 47
3.17 Time of Flight Upstream and Downstream detectors. Photos by A.
Norrick . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
3.18 A diagram of the Cerenkov detector used at MTest (Fermilab).
Copyright Fermilab. . . . . . . . . . . . . . . . . . . . . . . . . . . 49
3.19 Spectra of de Cerenkov detectors at MTest for the MINERvA Ex-
periment. Work done by M- Ramirez . . . . . . . . . . . . . . . . . 50
3.20 Underground trigger timing with the MINERvA Main Detector. . . . 51
3.21 Underground Trigger timing within the Test Beam Detector. . . . . 52
3.22 Logic of the triggers in the Auxiliary Test Beam detectors. . . . . . 53
4.1 Fermilab Accelerator Complex. . . . . . . . . . . . . . . . . . . . . 58

4.2 A niobium-based 1.3 GHz nine-cell superconducting radio frequency
to be used at the main LINAC of the International Linear Collider.
Photo from FNAL. . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
4.3 A pillbox cavity. The lower mode frequency does not depend from
the height of the cavity. . . . . . . . . . . . . . . . . . . . . . . . . . 62
4.4 The resonance curve’s full width is equal to the central frequency
ω0 dived by Q. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
4.5 Graphic description of a Bucket and a Bunch, the RF voltage that
the particles see while they are in a bunch and how the buckets
arrange themself on a synchrotron’s ring. . . . . . . . . . . . . . . 67
4.6 How is the booster bacth formed and the value of it. . . . . . . . . 69
LIST OF FIGURES xii
4.7 Diagram showing the MI’s cycle formation, spill duration and spill
frequency. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
5.1 A standardd histogram with box plots generate in ROOT. . . . . . . 73

5.2 Structure of the data in a ROOT file. . . . . . . . . . . . . . . . . . 74
5.3 Fitting a histogram made with three lines of code. . . . . . . . . . . 75
5.4 Structure of the data in a ROOT file. The comments below the
plots are the ROOT commands. . . . . . . . . . . . . . . . . . . . 77
5.5 Diagram of how the data is ordered in the ROOT file. The variable
shown is Time of readout. (a) and (c) are the energies that Run 1
contains for electrons and pions in both polarities (all the energies
are presented in tab. 5.1). (b) Run 2 and Run 3 contain onlu
electrons and pions. (d) Each Run has a number of subruns. (e)
Each subrun constains the the data for 1000 gates or events. (f)
One of the variables is the T ime variable as shown in this figure. . 80
5.6 Bias in the time profile by using an internal reference point mea-
sure in seconds. As can be seen, the (b) elecction allow us to
reduce the lost of events between the zero point and the first event. 82
5.7 Distribution of the values of the interval between the first event
recorded (kickoff od the spill) and two reference points: the $39
signal and the beginning of the subrun. . . . . . . . . . . . . . . . 83
5.8 File that contains dates from the signal $39 as a reference point. . 84
5.9 Number of spills for Run 2 and Run 3. . . . . . . . . . . . . . . . . 85
5.10 MI’s Spill frequency during May 2015. Infor provided by A.D. . . . . 86
5.11 How the TbTaTool produced the time profile. In (a) we show a
diagram of how the spills come in time and are arranged by the
variable Spilln umber. As it is shown, the point (2) and (3) were
used as a reference point. The TAT ”cut” the spills (b) indepen-
dently from one another and then, stack them in one plot (d). The
plot can be produced for different energies, polarities or type of
particles (c). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
LIST OF FIGURES xiii
5.12 From (a) to (e) apply to one subrun. (b) the different points of
reference in order to get the time profile. (c) and (d) explain how
the spill duration and the spill frequency are calculated. When the
calculation for one subrun is done, the next is calculated until the
last sunrun. (f) and (g) describe how all of them are stacked and
plot (i) and (j) according to different variables (h). . . . . . . . . . . 88
6.1 Time profile for Run2 ±4, ±6 and ±8 GeV (pions). . . . . . . . . . 90

6.2 Time profile for Run2 ±9, ±10 and 16GeV (pions). . . . . . . . . . 91
6.3 Time profile for Run3 ±2, ±4, ±8, 3 and 5GeV (electrons) . . . . . 92
6.4 Spill frequency for Pions (a,b) and Electrons (c,d). Pions have en-
ergies of 4, 6, 8, -8, 9, -9, 10 and 16 GeV, and electrons 2, -2, 4,
-4, 8, -8, 3 and 5 GeV . . . . . . . . . . . . . . . . . . . . . . . . . 93
6.5 Spill duration for all Runs (a,b) and Electrons (c,d). Pions have
energies of 4, 6, 8, -8, 9, -9, 10 and 16 GeV, and electrons 2, -2, 4,
-4, 8, -8, 3 and 5 GeV . . . . . . . . . . . . . . . . . . . . . . . . . 94
6.6 Errors against the oficial values of the MI’s frequency spill and MI’s
spill duration. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
A.1 Spill frequency for 4GeV Pions . . . . . . . . . . . . . . . . . . . . 100

A.5 Spill frequency for 10GeV Pions . . . . . . . . . . . . . . . . . . . 101
A.6 Spill frequency for 10GeV Pions and all the events . . . . . . . . . 102
A.7 Spill frequency for 2GeV electrons . . . . . . . . . . . . . . . . . . 103
A.8 Spill frequency for 4GeV electrons . . . . . . . . . . . . . . . . . . 103
A.9 Spill frequency for 3GeV and 5GeV electrons . . . . . . . . . . . . 103
A.10 Spill frequency for (±)8GeV and all energies electrons. . . . . . . . 104
A.11 Spill duration for data set containing only pions (Run2). . . . . . . 105
A.14 Spill duration for electrons (Run3). . . . . . . . . . . . . . . . . . . 107
LIST OF FIGURES xiv

B.1 (a) Left. Run Control of the MINERvA Main Detector. (b) Right.
Near MINOS DAQ, not all information is useful for MINERvA, just
the two bottom columns.) . . . . . . . . . . . . . . . . . . . . . . . 110
B.2 (a) Up left. NuMI Beamline Status Display which shows the status
of the beam. (b) Up right. MINERvA Veto Wall control. (c) IF
Beam Data Server A9 event monitoring. (d) Bottom left. MINERvA
Quality Checklist. (e) Bottom right. Live event display for neutrino
interactions in the Main Detector. . . . . . . . . . . . . . . . . . . . 111
B.3 Diagram of the MINERvA DAQ System. The Detector Expert train-
ing include a deeper knowledge in these areas. . . . . . . . . . . . 112
B.4 Cross Talk tasks . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
B.5 (a) Left. Some photos the Sillicon Detector Facility and the equip-
ment where the PMTs’ tests were done. (b) Right. Run and Slow
Control and some histograms which confirmed the presence of
cross talk. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
B.6 Monte Carlo simulated Rock Muon interaction with the detector. . . 115
B.7 Real interaction of Rock Muons with the detector. . . . . . . . . . . 116
C.1 Examples of wave guides. . . . . . . . . . . . . . . . . . . . . . . . 123

C.2 A pillbox cavity. The lower mode frequency does not depend from
the height of the cavity. . . . . . . . . . . . . . . . . . . . . . . . . . 127
C.3 The resonance curve’s full width is equal to the central frequency
ω0 dived by Q. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
List of Tables
1.1 Range of interaction of the four fundamental forces. . . . . . . . . 6

1.2 Sensitivity of different oscillation experiments.[32, p. 11] . . . . . . 12
2.1 Experiments focus on cross-sections and oscillation of neutrinos

and the type of channels of interaction focus on. . . . . . . . . . . 25
2.2 Neutrino experiments and their localizations. . . . . . . . . . . . . 25
2.3 Experiments and their detection technology. . . . . . . . . . . . . . 26
3.1 Modes of operation of the CAMAC 3377 TDC. The ? means that
there is no information . . . . . . . . . . . . . . . . . . . . . . . . . 46
3.2 Lowest momenta value for detection in Cerenkov detectors at MTest
(Fermilab) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
4.1 Time structure of the Beam according to Accelerator Division. . . . 67
5.1 Configurations of the Test Beam detector and the type of particles
that contain them. . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
5.2 Number of events for Run2 and Run 3 considering all the spills,
equal to 10, 6 or 2 spills per subrun. . . . . . . . . . . . . . . . . . 84
D.1 Calculation of the number of events in one subrun. . . . . . . . . . . . 132

D.2 Calculation of number of events for a set of root files. . . . . . . . . . . 133
D.3 Getting the unix timestamp for one subrun. . . . . . . . . . . . . . . . 133
D.4 Getting the unix timestamp for the beginning all the spills in one subrun. 133
D.5 Return one timestamp for the real beginning of one subrun. . . . . . . . 134
D.6 Function that create a txt file with beginning timestamps of root files. . . 134
D.7 Function that matches a set of two points. . . . . . . . . . . . . . . . . 135
xv
LIST OF TABLES xvi
D.8 Function that matches a set of two points for various root files. . . . . . 135
D.9 Function that generate the plots that shows the interval between all the
matched points. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
D.10 Function that plots the histograms of one variable. . . . . . . . . . . . . 136
D.11 Plotting stacked histograms for all energies . . . . . . . . . . . . . . . 136
D.12 Plotting one variable for polarities . . . . . . . . . . . . . . . . . . . . 137
D.13 Generating histograms for one variable for all energies and stacked them
in one plot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
D.14 Match of to points. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
LIST OF TABLES xvii
Introduction
In Chapter 1, a theoretical review of weak interactions is done in order to
describe the weak sector of the Standard Model (sec. 1.1) and introduce the
neutrino oscillation research (sec. 1.2). By now we know that the deficit in the
solar and atmospheric neutrinos reaching the earth is due to neutrino non-zero
mass which express in neutrino flavour oscillations. However in order to confirm
the value of masses between the three neutrinos more precise experiments are
needed in the reduction of systematic erros due to neutrino interactions with the
detectors. This is the motivation of the neutrino experiments like MINERvA Chap-
ter 2 and 3. The improve of detection models between neutrinos and nucleus,
since modern experiments are and will rely more on heavy nucleus detectors like
argon or lead. The calorimetric response and the fine detection of last product pi-
ons by the detector allows to reconstruct the incoming neutrino energy. MINERvA
face this challenge by setting up a small scale replica detector called Test Beam
Detector Chapter 4 where we receive particles of knwo momenta and type in
order to improve the models detection of pions, muons and electrons of the main
detector.
But, as important of knowing the composition of the beam, it is important to
know the time structure of it, in order to set up timing resolution of your detection
system. That is what this analysis has carried out, a comprobation of the timing
in the scale of Main Injector time space: the spill. In Chapter 4 and Appendix
C a review of the Time Structure of the Beam and the Electromagnetic theory of
Radio Frequency Cavities are stated. Section 4.3 describe why the beam that we
received have a specific structure in time.
In Chapter 6, we describe the Analysis Tool developed in C++ and ROOT that
allow us to analysis the scale of spill. Its features and how the tool works close
that chapter. Finally in Chapter 7, Appendix A and Chapter 8 we present the
results and conclusions fo this work.
I must mention that I had a small contribution in two publications of MINERvA:
(a) Physical Review D Vol 92 (2015) 9, 092008
LIST OF TABLES 1
Tı́tulo: Charge pion production in interactions on hydrocarbon at <E>!=

4.0GeV
Autores: G. Salazar, A. Zegarra, C. J. Solano Salinas and the MINERvA Colab-
oration
(b) Physical Review Letters Vol 116 (2016) 081802

Title: Measurement of Electron Neutrino Quasielastic and Quasielasti-clike
Scattering on Hydrocarbon at <E>!= 3.6GeV
Authors: G. Salazar, A. Zegarra, C. J. Solano Salinas and the MINERvA colab-
oration)
Finally in Appendix B, a review of other activities during the intership at Fermi-

lab has been described. In the Appendix D we present the documentantion and
in Appendix E the parts of the code. The final tool can be found in a public
repository at github1 .
1
https://github.com/gsalazarq/TbTaTool/tree/master/TbTaTool
Chapter 1
Theoretical Framework
One of the greatest achievements in science, is the formulation of the Standard

Model of Particle Physics, a quantum field theoretical framework which has the
most complete description of the fundamental process in nature.
As a quantum field theory, the Standard Model (SM) has been formulated in
terms of Lagrangians, with three fields: Gauge fields (bosons of spin 1 whose
mediate forces), Weyl fermions (which outline massless neutrinos) and a spin 0
scalar field which describe the Higgs boson. These SU(3)xSU(2)xU(1) gauge
fields portray what types of particles and interactions are allowed, where the
SU(3) gives rise to the strong interactions, SU(2) and U(1) describe the weak
and electromagnetic interactions.
The success of this model, has been stated because the predictions of new
phenomena and particles in a very precise way while the technology started to
reach the values of energies required. For example: in 1974, the discovery of
the J/psi (c quark); in 1977, the b quark; in 1981/82, W+- and Z bosons were
discovered. The appear of the Higgs boson in 2012 is probably one of the most
important milestones achieve by the Standard Model.
But, as impressive the predictions of the SM are, also the gaps and new find-
ings that do not fit in the theoretical framework. The baryon asymmetry of the
Universe ([23, p. 9]), the identity of dark matter, for how long the proton lives? or
the neutrino’s masses are a couple of flags inside the model. It is the neutrino
oscillations and the related neutrino cross-section scattering research that give
2
CHAPTER 1. THEORETICAL FRAMEWORK 3
experiments like MINERνA the fuel to look deeper into the nature of interactions
of neutrinos with matter.
One of the most impressive feature of the neutrino is that is not massless,
in contradiction with the Standard Model. Neutrinos research is one of the most
interesting, growing and intensive area of research in particle physics, with fu-
ture experiments that involve billions of investment as is the case of the DUNE
experiment.1 2
It is worth to mention that the 2015 Physics’ Nobel Prize went to T. Kajita
and A. McDonald for the neutrino’s oscillations research in Super-Kamiokande
(Japan) and the SNO Experiment (Canada) (fig. (1.1)):
”For the discovery of neutrino oscillation which shows that neutrinos

have mass”3
Other neutrino Nobel Prizes have been awarded in 1988 to L. M. Lederman,

M. Schwartz and J. Steinberger ”for the neutrino beam method and the demon-
stration of the doublet structure of the leptons through the discovery of the muon
neutrino”; in 1995 to M. L. Perl ”for the discovery of the tau lepton” and F. Reines
”for the detection of the neutrino”; in 2002 to R. Davis Jr. and M. Koshiba ”for
pioneering contributions to astrophysics, in particular for the detection of cosmic
neutrinos” and R. ”for pioneering contributions to astrophysics, which have led to
the discovery of cosmic X-ray sources”4 . For sure, those Nobel Prizes will not be
the last one to be awarded to neutrino research.
1.1 Weak Interactions
Decays of µ and τ decays’ and natural radioactivity are due to weak interactions.
Even that unified weak and electromagnetic theory was proposed by S. Weinberg
1
The Deep Underground Neutrino Experiment (DUNE) is a proposed experiment with a near
detector at Fermilab and a far detector at the Sanford Underground Research Facility at South
Dakota. This international mega-science project is designed to discover, for example, if neutrinos
exhibit matter-antimatter asymmetries. More information at http://www.dunescience.org/
2
Ghosh, Pallab (15 February 2014). ”UK backs huge US neutrino plan”. BBC News. Retrieved
15 February 2014.
3
http://www.nobelprize.org/nobel_prizes/physics/laureates/2015/
4
The quotations have been reviewed from the official page of the Nobel Prize Foundation
http://www.nobelprize.org/nobel_prizes/physics/laureates/
Figure 1.1: The 2015 Nobel Prize in Physics winners: Takaaki Kajita (left) and
Arthur B. McDonald (right)
and A. Salam in the late sixties, the actual discovery of the mediators did not
happen until January of 1983.
During the decade 1964-1974 ([16, p. 42]), the particle physics theoretical
framework was incomplete and full of new discoveries that did not fit on it. In the
summer of 1974, the ψ meson was first observed at Brookhaven Laboratory by
a group under C.C. Ting, with a lifetime of 1000 times greater than any particle;
by next year it was a new lepton: the tau[30]. In 1983, the W was discovered by
Carlo Rubbia’s group at CERN5 (at 81 ± 5GeV /c2 and five months after, the Z 0 (at
95 ± 3GeV /c2 )6 .
Now we know that matter is made out of three kind of elementary particles:
leptons, quarks and mediators. There are three generations of leptons, according
to their charge (Q), electron number (Le ), muon number (Lµ ) and tau number (Lτ ),
each of these numbers define a generation or a family, composed by a lepton and
its corresponding neutrino. The classification ends considering the particles and
antiparticles, the weak force having two mediators for charge currents (W ± ) and
one for neutral current (Z 0 ).
These three mediators (W ± , Z 0 ) correspond to the triplet for SU (2)L and a
single, for U (1)Y The symmetry breaking in the SU (2)L × U (1)Y sector gives the
5
G. Arnison, Physics Letters B, Volume 122, Issue 1, 1983, Pages 103-116.
6
G. Arnison, UA1 Collaboration. Physics Letters B, Volume 126, Issue 5, 1983, Pages 398-
410.
three SU(2) mediators mass thought the interactions with the Higgs Boson. The
photon, though stay massless.
Since the three mediators have positive, negative and no charge, the inter-
actions can be ordered into charged currents interactions and neutral currents
interactions.
The fundamental charged vertex is presented in the fig. (1.2). A neutrino is
produced by its corresponding lepton, with the emission of a W − (or absorption
of W + )
Figure 1.2: Fundamental vertex for weak interactions.
A more complicated reactions can be produce if the primitive diagram is com-

bined, as for example for the reactions:
µ− + νe → e− + νµ
(1.1)
ν¯e + p → n + e+
If the target is a nucleon and with the necessary energy of the income neu-
trino, we can resolve the nucleus as a one (CCQE - Quasielastic Interactions) or
as a quarks (DIS - Deep Inelastic Scattering). For example, the reaction (1.1)
is called inverse neutron decay and was used by C. Reins and F. Cowan in the
discovery of the antineutrino. Also, weak charged current can change lepton and
quark flavours.The fundamental neutral current (fig.2.2) was first suggested in
1958 by Bludman7 , that preserves the three leptonic numbers.
The 1973 CERN’s bubble chamber, revealed that in the reaction ν¯µ + e− →
ν¯µ + e− , a neutral Z 0 was the mediator8 . The same experiment saw the mediation
7
C. D. Anderson. ”Early Work on the Positron and Muon ” American Journal of Physics De-
cember 1961 Volume 29, Issue 12, pp. 825
8
DESY found unmistakable evidence of the contribution of Z 0 by studied the reaction e− +
of this boson in neutrino-nucleon scattering (1.2), with values of three times large
as those related with charged events [16, p. 323].
νµ + N → νµ + N (1.2)
From the values of decay lifetimes we can be inferred the forces that produces
and the strength of the interaction. Pions and muons decays’ lifetimes (1.3) are
considerably longer than particles that decay only by strong and electromagnetic
forces. The difference in the order of magnitudes are evidence of the existence
of another type of interaction, beside the strong interaction.
π − → µ− ν¯µ , with τ = 2.6 × 10−8 sec,

µ− →− ν¯e νµ , with τ = 2.2 × 10−8 sec, (1.3)
Finally we recall that the lifetimes are inversely related to the coupling strength
of this interactions, that means that this new interaction is has weaker coupling
than electromagnetism.
Table 1.1: Range of interaction of the four fundamental forces.
Interaction Range Lifetime Cross Section Coupling

(sec) (mb) αi
Strong 1F ' 1
mπ 10−23 10 1
Color
confinement e.g. ∆ → pπ e.g. πp → πp
range
Electromagnetic ∞ 10−20 ∼ 10−16 10−3 10−2
e.g., π 0 → γγ e.g., γp → pπ 0
Σ → Λγ
Weak 1
MW 10−12 or 10−11 10−6
with longer
MW ' 100mp e.g., Σ− → nπ − e.g., νp → νp
π − → µ− ν̄ νp → µ− pπ +
e+ → µ− + µ+
1.2 Neutrino Physics
Neutrino was proposed by Wolfgang Pauli in 1930 to hold up the energy-momentum

conservation law and Fermi’s statistics in β-nuclei decay (eq. 1.4). This new par-
ticle had to have: no charge, should not interact with matter, or in the other case
interact very weakly and almost no mass. For the same problem Bohr, proposed
a statistical version of the energy conservation law. Experiments showed that
Pauli was right.
n0 → p+ + e− + ν¯e (1.4)
In 1932, a more massive neutral particle was discovered by Chadwick([8]),

and was named neutron. Pauli’s name of the unknown particle was turned into
neutrinos (little neutron one) by Enrico Fermi, who used it in 1932.
In 1956, the antineutrino was discovered experimentally by F. Reines an C.L.
Cowan allowing the interaction of protons with electron antineutrinos’ fluxes9 cre-
ated in nuclear reactions (eq. 1.5):
ν¯e + p+ → n0 + e+ (1.5)
The detection was made since a positron quickly finds an electron producing
two opposite 0.5 MeV γ rays which are detectable, but not necessarily the indica-
tion of the antineutrino existence. Another gamma ray is detected due to capture
of the neutron by the nucleus (n +108 Cd →109 Cd∗ →109 Cd + γ). This two events
configure the signature of an antineutrino interaction looked.
Neutrinos are elementary particles with spin 1/2, electrically neutral and obey
Fermi-Dirac statistics. The Standard Model consider three Weyl massless neu-
trino flavors: νe , νµ , ντ , each one corresponding to three different leptons, the
electron e− , the muon µ− and the tau τ , each doublet has their antiparticles.
However the nature of neutrino is still open, for sure they are not Weyl’s mass-
less fermions. Two plus one options arise: neutrinos can be Majorana neutrinos,
9
By that time they were expecting 1012 − 1013 neutrinos per second per m2 . They did not know
that they were using electronic antineutrinos.
Figure 1.3: Schematic diagram of neutrino detector as appeared in the original

paper of Reines and Cowan10 .
Dirac neutrinos or a mix between both where the seesaw mechanics plays an
important role.
Parity and CP invariance
Parity is violated in weak interactions as Goldhaber observed, neutrinos have spin
antiparallel to their momentum (left-handed) and antineutrinos have it parallel
(right-handed).11
Not only the parity is violated, but also the charge conjugation invariance,
where Γ is the lifetime of process. For instance,
Γ(π + → µ+ νL ) 6= Γ(π + → µ+ νR ) P violation,

(1.6)
Γ(π + → µ+ νL ) 6= Γ(π − → µ− ν¯L ) C violation,
but the CP variance is conserved. Future experiments have as main objective

the comprobation of the CP invariance or its violation by neutrinos.
Γ(π + → µ+ νL ) = Γ(π − → µ− ν¯R ) CP invariance. (1.7)
1.2.1 Solar and Atmospheric Neutrinos
The neutrino mass, is one of the most important discoveries from the last decade
outside the framework of the Standard Model12 .
11
The historic experiment for testing this, was with β-transitions of polarized cobalt nuclei.
60
C →60 N i∗ + e− + ν¯e
12
A good review of the physics beyond the Standard Model is made by N. Mihoko in [24]
Solar Neutrinos
The Solar Neutrinos problem was first formulated in 1964 by Ray Davis’s and
John N. Bahcall from the Homestake Experiment13 , who were the first to look
a deficit in the flux of neutrinos from the Sun. Their final results, published in
1998[10], showed that the experimental value, 2.56 ± 0.16 (stat) ± 0.16 (sys) was
over 30% of the theoretically flux value 8.5 ± 0.9 SNU [28].14
So why the experimental flux that reaches the earth is anomalously low? This
is the core question of the Solar Neutrino Problem[1, p. 10].
We are able to distinguish the solar neutrinos, since we know the sun’s pro-
cesses that generate them: fusion of hydrogen to helium: p + p → 2H + e+ + νe ,
the pp-chain 4p →4 He + 2e+ + 2νe and the CNO-cycle are process that produce
neutrinos[21, p. 9-7].
While measurements of the different solar neutrinos’ chains were improved,
in the flux did not change. The Sudbury Neutrino Observatory (SNO) with a
Cerenkow detector of 1000ton ultra-pure heavy water (D2 O) in an acrylic sphere
of 12 m diameter, clearly state that the deficit was not a technical problem, in-
stead the conversion between νe and νµ was a physical event, later confirmed by
KamLAND reactor experiment.15
Atmospheric neutrinos
As we know, cosmic rays from outer space interact with the atmosphere generat-
ing particles that come into the earth’s surface. Through these decays processes
(eq. 1.8), we are able to study the flux atmospheric of electron and muon neu-
trinos putting the detectors underground, shielding them from the muons and
electrons generated.
13
A lot of information about the experiment and the papers can be found on the web page
http://www.nu.to.infn.it/exp/all/homestake/
14
A SNU is the solar neutrino unit (SNU). It is equal to the neutrino flux producing 10−36 cap-
tures per target atom per second.
15
As mentioned before, with the Super-Kamiokande Collaboration won the 2015 Physics Nobel
Prize.
π ± → µ± + νµ (ν¯µ )
(1.8)
µ± → e± + νe (ν¯e ) + ν¯µ (νµ )
At low energies (≤ 1GeV ), the ratio between the flux of muon-neutrinos to

electron-neutrinos was around ∼ 2[15]16 . A better ration index was later improved
defining R as the ratio of data to theoretical expectation fluxes. The IMB experi-
ment 17 reported R ∼ 0.54[19] and KamiokandeR ∼ 0.60[18].
Later, T. Kajita from Super-Kamiokande experiment18 presented compelling
evidence in favour of neutrino oscillations in the neutrino conference Neutrino’98[14].
(a) (b)
Figure 1.4: (a) Path of the atmospheric neutrinos during its travels through the
earth. (b) Zenith angle events distribution of e-like and µ-like events in Super-
Kamiokande in the range of energy below 1.33GeV[11]. The red line, indicates
the best fit for the data points, while the boxes show the Monte Carlo events
expectation considering no oscillations.
If the flux of neutrinos coming from the atmosphere is expected to be isotropic,

independent of the zenith, a question raised, why the observed fluxes of up-going
16
The ratio is defined as R = (Nµ /Ne )DAT A /(Nµ /Ne )M C
17
The Irvine-Michigan-Brookhaven Water-Cerenkov detectors (Ohio, USA) http:
//www-personal.umich.edu/~jcv/imb/imb.html
18
Super-Kamiokande Collaboration a 50000t water Cerenkov detector, 10 times larger than its
predecessor Kamiokande in the Mozimu zinc mine, Japan.
and down-going neutrinos in an underground detector are not the same?. While
the flux of e-neutrinos has almost no zenith angle dependece (right plots from fig.
1.4), the µ-neutrinos’s flux of down-going (cosθ = 1) exceeds the flux of up-going
νµ .
The most simple explanation was accepted: neutrino flavour oscillation. Neu-
trinos moving upward through the detector are created in the atmosphere at the
opposite side of the Earth, and travel thousand of kilometers before interaction.
Apparently muon-neutrinos disappear on the way whereas electron-neutrinos
do not. Down-going muon-neutrinos, produced in the atmosphere directly above
the detector, only travel a few dozen kilometers and are detected at the level
expected. No indication of an increased electron-neutrino flux, the missing muon-
neutrinos must have oscillated into tau-neutrinos.
After 1998, neutrino oscillations started to open a new set of question that
needed to be answered through the design of new experiments focus on neutrino
oscillations, and as we will mention, a shears experiments started to work in the
improvement of the models of neutrino-nucleus interaction, like MINERνA (sec.
3.1). In the next section, a brief review of the oscillation of only two flavours are
made.
1.2.2 Neutrino oscillations
Consider for simplicity that two of the known neutrinos νe and νµ are eigenstates
with no well defined masses but they are a linear composition of the neutrino
mass eigenstates ν1 and ν2 with masses m1 and m2 , respectively:
|νe i = |ν1 icosθ + |ν2 isinθ

(1.9)
|νµ i = −|ν1 isinθ + |ν2 icosθ,
where θ is the neutrino mixing angle. We will get the dependece of the oscilla-
tion probability with the energy and the difference of masses. Following the rules
of quantum mechanics, we can construct the state at any time t, and then get the
probability of transformation from a state to another.
First, lets consider that at time t = 0. We have a pure weak eigenstate
|ν(0)i = |νµ i , the propagation of this state in time is dictated by the free non-
time dependent Hamiltonian.
|νt i = −|ν1 ie−iE1 t sinθ + |ν2 ie−iE2 t cosθ, (1.10)
m21,2
q
where E1,2 = p2 + m21,2 ≈ p + 2p
. The probability of finding a neutrino with
electron flavor is then:
P (νµ → νe ; t) = |hνe |ν(t)i|2

−iE2 t
= sin2 θcos2 θ| − e−E1 t+e |2
∆m2 t
= sin2 2θsin2 ( ) (1.11)
4E
∆m2 L
= sin2 2θsin2 ( )
4E
Here ∆m2 = m22 − m21 is the squared mass difference and E = p. The last
line is valid for relativistics particles (L = t) with L being the traveled distance,
which in practice allow us to construct tunels of decays in order to observe the
oscillation, or in turn place the far and near detector in an oscillation experiment
at a fixed distance.
As it can be seem the mass difference is present in the form of a squared
difference, hence the measuring oscillation probabilities will not give absolute
values of the masses. In the table (1.2), different sources of neutrinos and the
minimum value that can reach the measurement of min(∆m2 )[eV 2 ].
This model with two flavors, can be extended to one with three flavor mixing
and three angles, three squared massess differences ∆m212 , ∆m213 and ∆m223 .
[20]
Table 1.2: Sensitivity of different oscillation experiments.[32, p. 11]

Source Type of ν E[MeV] L[km] min(∆m2 )[eV 2 ]
Reactor ν¯e ∼1 1 ∼ 10−3

Reactor ν¯e ∼1 100 ∼ 10−5
Accelerator νmu
¯ , νµ ∼ 103 1 ∼1
Accelerator νmu
¯ , νµ ∼ 103 1000 ∼ 10−3
Atmospheric ν’s’ νe,µ
¯ , νµ,e
¯ ∼ 103 104 ∼ 10−4
Sun νe ∼1 1.5 × 108 ∼ 10−11
Chapter 2
Neutrino Experiments
In recent years, many experiments1 have been setup in order to improve the mod-
els of interaction between detectors and neutrinos due to requirement of neutrino
oscillation experiments.
Neutrino interaction’s models need to predict not only the signal and back-
ground of mass neutrino oscillation, but also how the energy is transfered to the
observable particles, while reducing the uncertainties2 . As mentioned by D. Har-
ris: [17, p- 1]
”Future oscillation experiments such as DUNE [9] depend on the ability to
predict far detector signal’s (background) spectra at the 1% (5%) level (...), the
particle physics community is still at the level of measuring cross sections and
making far detector predictions at the 7 − 10% level.”
New oscillation experiments need to expand the models for other nucleus
than hydrogen or deuterium, since the far and near detectors are made of targets
of carbon, water, argon or iron. Another requirement for neutrino cross-section
experiments is to focus energy region of few hundred MeV to a hand full of GeV,
since oscillation probabilities are function of the inverse of the neutrino energy.
It is worth to mention that MINERνA has a 15% constrain from a CCQE mea-
surements and 10% in the flux uncertainties ”[17].
1
Only for mention some of them: MINERνA, T2K (Tokai to Kamioka), NOνA, MINOS and
MicroBooNE
2
Another source of uncertainty comes from the incoming neutrino flux which is now at 8% to
10%.
14
CHAPTER 2. NEUTRINO EXPERIMENTS 15
MINERνA3 is a cross-sections precision studies experiment of (anti)neutrino-

nucleus scattering in the range of 1-20GeV at the NuMI Beam at Fermilab. The
tracking region is made purely by scintillators which recognize the particle by
the energy loss per unit length (dE/dx) after the neutrinos had interacted with
the targets (carbon, iron and lead) interleaved between the scintillator planes.
Technical details of the experiment will be describe in sec. 3.1.
This chapter state the principal concepts of neutrino interactions with mat-
ter (sec. 2.1), the different regions of cross-section interaction ν-A regarding
the energy (sec. 2.2) and experiments that perform measurements of neutrino’s
properties (sec. 2.3).
2.1 Neutrino interaction with matter
The cross-section (σ) quantify the interactions between particles and targets, re-
garding energy, flux and type of interactions taking place (if it is a strong, electro-
magnetic or a weak interaction).
Defined as the rate (sec. 2.1) of interactions between incoming particles that
are scattered due to the interaction with the targets over a known energy and flux
of incoming particles,
Number of reactions of a given type per unit time

σ= (2.1)
(Incoming flux)(Number of target particles)
σ is a number that has dimensions of area and is usually expressed in cm2 or

in barns (1barn = 10−24 cm2 )45 .
Differential cross-sections (dσ/dA) is a distribution of probability, which give
us the dependence of the cross-sections to an specific variable (A), for example
the angle range dθ around some direction θ.
An elastic cross-sections, is a type of interaction where neither beam parti-
cle or the target has been disintegrated, the opposite is called an inelastic cross-
3
Main INjector ExpeRiment ν-A
4
The proton-proton cross-sections is 40mb
5
For each cross-section corresponding to a particular type of interaction, the term partial is
added.
Figure 2.1: Total cross-sections for Charge Current Quasi-elastic Scattering on

Carbon measured by different experiments. This figure has been taken from the
Conference MINERνA 101 by C. Patrick [27]
Figure 2.2: Fundamental vertex. This figure has been taken from the Conference
MINERνA 101 by C. Patrick [27]
sections. When we sum the inelastic and elastic interaction, the cross-sections
is called total cross-sections, while the inclusive cross-sections is referred to
all the process that contains at least one π + in the final state (e.g., p + p → π + )
an exclusive cross-section is when the final stated is exclusively defined with
no extras (e.g., p + p → p + p + π 0 ).
In the next section, it will be described the three most important process of
this intermediate energy region6 .
2.1.1 Intermediate Energy Cross Sections
In the range of Eν ∼ 0.1−20GeV three main categories in neutrino scattering are:

elastic and quasi-elastic, resonance production and deep inelastic scattering.
The elastic and quasi-elastic scattering is produce when a neutrino elas-
tically scatter off a nucleon target, liberating a nucleon (or many of them) from
it. Quasi-elastic scattering is also referred as charged current neutrino scattering
(CCQE), while neutral current scattering is traditional named as elastic scattering.
The next region is the resonance production region going up in the range
of energy, neutrinos can excite nucleon’s target into a baryonic resonance state
(∆, N ∗ ), and decay into many mesonic final states producing combinations of
nucleons and mesons.
Finally, the deep inelastic scattering’s energy allows the neutrino resolve
the individual quark constituents of the nucleon, that usually is expressed with
the creation of hadronic showers. Nuclear effects have more impact in the cross-
section scattering in the subregion. The fig. (2.3) summarizes the total neutrino
and antineutrino per nucleon in the CC Cross Section (charged current interac-
tions).
6
The complete classification made by J. A. Formaggio and G.P. Zeller [13] is: (1) Threshold-
less processes (Eν ∼ 0 − 1MeV), (2) Low Energy Nuclear Processes (Eν ∼ 1 − 100MeV), (3)
Intermediate Energy Cross Section (Eν ∼ 0.1 − 20GeV), (4) High Energy Cross Section (Eν ∼
20 − 500GeV) and (5) Ultra High Energy Neutrinos (Eν ∼ 0.5TeV −1EeV)
Figure 2.3: Total neutrino and antineutrino per nucleon CC cross-sections divided
by neutrino energy and plotted as a function of energy. Data include the low
energy data from N ([6]), ∗ ([]), ([]), and ? ([]). The three main regions are ploted
as quasi-elastic (dashed), resonance production (dot-dash), and deep inelastic
scattering (dotted). The predicitions for each region are provided by the NUANCE
generator ([7])
Quasi-elastic Scattering
Quasi-elastic (QE) interactions are produced when a neutrino scatter off a nu-
cleon and create a charged lepton. A typical reaction (eq. 2.2) produce a proton,
a neutron and the corresponding charged muon. If the energy of the neutrino
is less ∼ 2GeV, is more likely to have a CCQE event. These events are the
dominant signal mechanics for T2K and a large fractions for NOvA experiment.
νµ + n → µ− + p,
(2.2)
ν¯µ + p → µ+ + n
For quasielastic interactions, the cross section is given by the Llewellyn Smith
formalism7
dσ G2F M 2 |Vud |2 (s − u) (s − u)2

= [A ± + C], (2.3)
dQ2 82ν M2 M4
where ± referes to (anti)neutrino scattering, GF is Fermi’s coupling constant,
Q2 is the squared four-momentum transfer (Q2 = −q 2 > 0) M is the nucleon
mass, m is the lepton mass, Eν is the incident neutrino energy, and (s − u) =
4M Eν − Q2 − m2 . The factors A, B and C are functions of the familiar vector (F1
and F2 ), axial-vector (FA ), Vud is element of CKM-Matrix8 for quark mixing, and
the pseudoscalar (FP ).
With CCQE interactions we can study the weak nucleon form-factors and fully
reconstruct the incoming energy neutrino9 which is critical for measuring the os-
cillation parameters. However, the uses for CCQE are complicated with heavier
targets.
It can be seen from the fig. (2.4) that the predicted values are bellow the
experimental data, this comes from the fact that nucleon-nucleon correlations
and two-body exchange currents must be included for improving the accuracy of
7
C.H. Llewellyn Smith, Phys. Rep. 3, 261 (1972).
8
N. Cabibbo (1963). ”Unitary Symmetry and Leptonic Decays”. Physical Review Letters 10
(12): 531–533 M. Kobayashi, T. Maskawa; Maskawa (1973).
”CP-Violation in the Renormalizable Theory of Weak Interaction”. Progress of Theoretical Physics
49 (2): 652–657
9
The initial nucleon is at rest and by reconstructing the momentum transfer simply by measur-
ing the outgoing muon’s momentum and angle and assuming conservation of total momentum
and energy.
neutrino-nucleus QE scattering.[17].
Figure 2.4: All νµ quasi-elastic scattering cross-sections, νµ n → µ− p measure-

ments until the year 2002 by [7] as a function of neutrino energy for different
nuclear targets.
Neutral Charge Elastic Scattering
According with the Glashow-Weinberg-Salam theory, the weak interaction has

also a neutral boson Z 0 bosons, that are responsible for neutral current interac-
tions like:
ν¯µ + e− → ν¯µ + e−
νµ + e− → νµ + e− (2.4)
With a threshold for this events of 2.2MeV, the neutrino transfers some of its
energy and momentum to the target particle and scatter off a very forwarding
charged lepton within small angles (< 5). The expression eq. (2.3) for the differ-
ential cross-sections is also applied to neutral charge elastic scattering.
Resonance Production
With more Q2 available, the neutrino enters in the frontiers of inelastic scatterings
between the 0.5GeV < Eµ < 10GeV while the production of leptons will look
the same, the hadronic sector is going to be more fruitful, pushed into baryonic
resonance state involving N ∗ or ∆. Some process that can be found are (2.5):
νµ n → µ− ∆++ → µ− + π + p
νµ n → µ− ∆+ → µ− + π + n
νµ p → µ− pπ + ν¯µ p → µ+ pπ −
νµ p → µ− pπ 0 ν¯µ p → µ+ pπ 0 (2.5)
νµ n → µ− nπ + ν¯µ n → µ+ nπ −
νµ p → µ− pπ +
νµ n → µ− pπ 0
The resonances then decay back down to a nucleon, accompanied by a single

pion most of the times (which is called ”Resonance Single Pion Production”)
or into multi-pion, other mesonic (K, η, ρ), and photon final states10 (π 0 → 2γ).
Another channel of interaction is when a neutrino coherently scatter off the en-
tire nucleus and produce a very specific forward-scattered single-pion final state
for CC νµA → µ− Aπ + , ν̄µA → µ+ Aπ − and for NC νµA → νµAπ 0 ,ν̄µ Aπ 0 . In
the fig. 2.5 we show the historical measurements of νµ CC resonant single-pion
production.
Neutrinos can also coherently produce single pion final states, where a neu-
trino scatters off an entire nucleus coherently and produces a very forward-going
pion and transfers little or no energy to the nucleus (fig. 2.6). With no nuclear
recoil, a distinctly foward-scattered pion and low-Q2 interactions, coherent pro-
duction of pions is present in NC and CC interaction, across a broad energy
range, which means that is poorly-understood.
10
Photon production processes are an important background signal in νµ → νe appearance
neutrino oscillation experiments, because it can easily look like an electron.
Figure 2.5: Data measurements of νµ CC resonant single-pion production reported

by experiments with no additional corrections derived of nuclear targets or invari-
ant mass selections. The continuous curve has been generated by NUANCE.
[26].
νµ A → νµ Aπ 0 ν̄µ A → ν̄µ Aπ 0
(2.6)
− + + −
νµ A → µ Aπ ν̄µ A → µ Aπ
Deep Inelastic Scattering
A deep inelastic scattering is generated when via the exchange of a virtual W ±

or Z 0 , the neutrino scatters off a quark in the nucleon producing a lepton and a
hadronic system as final states becoming a tool for study QCD. A DIS interaction
can be describe in terms of inelastic, 4-momentum transfer Q2 and the Bjorken
scaling variable[13].
νµ N → µ − X ν̄N → µ+ X
(2.7)
νµN → νµX ν̄N → ν¯µ X
2.2 Neutrino Cross Sections Measurements
This section try to address the principal challenges that arise in trying to un-
derstand the complex interactions in the field of low energy neutrino interactions
Figure 2.6: Absolute coherent pion production measurements from a variety of

nuclear targets and samples for NC and CC data [7, p. 33].
and their principal results. From strong evidence of neutrino oscillations and the
existence of neutrinos mass exist from atmospheric neutrinos, solar neutrinos,
reactor experiments, and long-line base oscillation experiments, the explanation
of
(...) why the neutrino masses are so small and their mixing are large
often rely on physics at the GUT scale. One of the most popular ideas,
known as the See-Saw mechanism, coupled with CP violation in neu-
trinos produces leptogenesis, where a lepton matter/antimatter asym-
metry caused by the decay of heavy neutrinos are converted into a
baryon asymmetry and explains why today we live in a matter domi-
nated universe.[33]
The neutrino oscillation’s hints were discovered in experiments with natural

neutrinos and uncontrolled conditions, but a new generation of experiments are
taking the advantage of artificial generated neutrinos, mainly in long-baseline
experiments around the world. All experiments face similar problems, namely
uncertainties in the neutrino energy reconstruction, since the presence of the
neutrino is detected indirectly, mainly by product’s particles of the interactions of
neutrinos with target nucleus in the detectors.
Since there is no cleanly interaction neutrinos with a single quarks, the ex-
periments that look for experimental oscillation measurements rely in detectors
where the knowledge of how the neutrinos interact with nucleus is more important
than before.
Due to this neutrino cross sections and nuclear effects’ uncertainties, there
is an estimated of 20% to 50% in oscillations experiments[20, p. 10]. Therefore,
with large errors, make a precise determination of oscillation parameters is more
difficult and can not be considered.
In the case of MINERνA, the uncertainties in the knowledge of neutrino scat-
tering measurements comes from the level of knowledge of the beam parameters
and the cross-sections, as well with to uncertainties of the detector itself.
The knowledge of neutrino cross-sections along this energy region can be
summarized from experiments conducted in 1970’s and 1980’s using bubble cham-
ber and spark chamber detector technology, and then scintillator technology and
liquid argon technology. Also of importance is the electroweak parameters (sin2 θW )
and structure functions in the deep inelastic scattering region. The following ex-
periments are still taking data and will update the figure (2.3): ArgoNeuT, K2K,
MiniBooNE, MicroBooNE, MINERνA, MINOS, NOMAD, SciBooNE and T2K (fig.
2.7).
Figure 2.7: Charge Current Neutrino interactions in the intermediate region and
the experiments that cover those regions. From M. Martin presented in the Nu-
Fact15 [22].
2.3 Neutrino experiments
The new emphasis in neutrino scattering has increased due to the oscillations
neutrino experiments, where both charged current (CC) and neutral current (NC)
channels have been collected over many decades using a variety of targets,range
of energy, analysis techniques and detector technologies.
The detector technologies needed a renewed appreciation for nuclear effects
and the importance of improved neutrino flux calculations, performed in the low
and medium energy with an emphasis on inclusive, quasi-elastic, and single-pion
production processes. The table 2.1 shows a list of modern accelerator-based
neutrino experiments, the table 2.2 show the neutrino oscillation experiments and
in the table 2.3 the detection technology that is used by some experiments.
Table 2.1: Experiments focus on cross-sections and oscillation of neutrinos and

the type of channels of interaction focus on.
Experiment beam Eν neutrino run σν

GeV target(s) period publications
ArgoNeuT ν, ~ν 3.3 Ar 2009-2010 CC
ICARUS ν 20.0 Ar 2010 - present
K2K ν 1.3 CH, H2 O 2003-2004 QE π
MicroBooNE ν 0.8 Ar 2014 - present
MINERνA ν, ~ν 3.3, 5.6 He, C, O, Fe, Pb 2009 - present QE
MiniBooNE ν, ~ν 0.8 CH2 2002-2012 QE
π[15,16,17,18,19]
MINOS ν, ~ν 3.3, 6.5 Fe 2005 - present CC
NOMAD ν, ~ν 26.0 C 1995-1998 CC , QE, π
NOvA ν, ~ν 2.0 CH2 2010 - present
SciBooNE ν, ~ν 0.8 CH 2007-2008 CC , π
T2K ν, ~ν 0.85 CH, H2 O 201 - present CC
Table 2.2: Neutrino experiments and their localizations.

Experiments Range Eν Located Source
K2K 1.5 GeV Japan Low neutrino beam

sent from KEK (started 1999)
with a near detector at KEK site
MINOS 3 -12 GeV Fermilab at a distance of 730km
(USA) in the NuMI beamline
high-energy neutrino beam
CNGS 17GeV Gran Sasso 730 km away of
(CERN) Neutrino source at CERN
Table 2.3: Experiments and their detection technology.

Experiment Scope Technology
T2K Reconstruction of Water

neutrino energy spectrum Cherenkov
MINERνA Reconstruction of Calorimetric
neutrino energy spectrum detectors
MicroBooNE Identification and Water
MiniBooNE Reconstruction of Large hydrid
tau neutrinos tracking/emulsion detectors
2.3.1 T2K experiment
T2K (Tokai to Kamioka) is a long-baseline neutrino experiment in Japan that

search for oscillations from muon neutrinos to electron neutrinos, produce by the
intense beam of muon neutrinos with a energy of hEν i ∼ 0.6GeV J-PARC in the
center of Japan, and directed towards the Super-Kamiokande detector 295km
away.
The near detector (INGRID) is situated 280 meters from the target in the cen-
tre of the neutrino beam, at its objective is to check the direction and intensity
by daily basis of the neutrino beam. The far detector on the other hand, is lo-
cated 100 meters underground in western Japan and is a very large cylinder of
ultra-pure water. By detecting the production of muon and electron neutrinos,
T2K has seen almost 7 times more electron-neutrino events than if there were no
oscillations.
The physics goals of the experiment is to calculate the νµ to νe oscillation, the

value of mixing angle θ13 , precision measurements of oscillation parameters in
νµ disappearance and the search for sterile components in νµ disappearace in
neutral-current events.
(a) (b)
Figure 2.8: Geographical map of the T2K detection the east coast of Japan. Im-
ages from the ofical page of T2K experiment http://t2k-experiment.org/.
2.3.2 MiniBoone
Using the beam from Fermilab’s Booster accelerator in the energy region of
Eν = 0.5 − 1GeV a 800 tons of mineral oil and 1280 PMTs dectector, MiniBoone
was aimed to confirm the excess of electro neutrino events and support the neu-
trino oscillation interpretation of the LSND (Liquid Scintillator Neutrino Detector)
experiment νµ νe and ν¯µ ν¯e .
One of the open questions is if there are more neutrinos (”sterile” neutrinos)
that would interact only through gravity. The LSND experiment sets hints in this
direction, puzzling the neutrino community.
As in same case of MINERvA, MiniBooNE detector receive neutrinos pro-
duced in the decays of mesons and muons, looking for few oscillation to occur,
since the decay length is only 500 meter away. The experiment results finally rule
out a fourth sterile neutrino.
Another currently operating experiment, MicroBooNE, is measuring low en-
ergy neutrino cross sections and investigating the low energy excess events ob-
served by the MiniBooNE experiment. Also, this experiment is testing the future
construction of massive kiloton scale LArTPC detectors for future long-baseline
neutrino physics (DUNE).
(a) MiniBoonE Phototube Support Struc- (b) MicroBooNE detector at

ture. Fermilab
Figure 2.9: MiniBooNE and MicroBooNE experiments

Chapter 3
MINERvA a cross-section neutrino

experiment
As mentioned in the former sections (2.1), (2.2) and (2.3) the neutrino oscilla-
tion experiments need better models of interaction with the detectors. However,
the signal and backgrounds measurements present in the process that are the
poorly measured. MINERvA provide data that considerably improve the models
neutrino-nucleus scattering and thus reduce the systematic uncertainties in the
result from oscillation experiments.
Our experiment (fig. 3.1) is a fine-grained, fully active neutrino detector placed
in the NuMI neutrino beam (3.1.1) with a good high-rate studies of neutrino-
nucleus interactions and good resolution of final states using νµ and ν̂µ incident
of 1-20GeV.
In this chapter I will describe the most important components of the MINERvA
detector (3.1), why the experiment includes a scale-down replica of the detector
(3.2.1) and finally the main components of the MINERvA’s Test Beam Experiment
are stated in the last section (3.2).
29
CHAPTER 3. MINERVA A CROSS-SECTION NEUTRINO EXPERIMENT 30
Figure 3.1: Front view of the MINERvA detector at NuMI Hall at Fermilab.
3.1 ”Bringing Neutrinos into Sharp Focus”
MINERvA1 is a neutrino experiment dedicated to explore precision neutrino cross-

section measurements in multiple nuclear targets in order to study nuclei and
nuclear effects, as mentioned in [33]:
The experiment provides the opportunity for a broad array of physics

studies, using neutrinos as probes to study nuclear processes and nu-
cleon structure as well as exploring the properties of neutrinos them-
selves. (..) Uncertainties in the cross sections for these processes
contribute to the systematic error of oscillation measurements, so im-
proved measurements of the cross section will contribute directly to
improved precision in the measurement of neutrino oscillation param-
eters.
By receiving around 1020 proton on target, the number of events for CCQE,
Coherent Production of Pions and DIS has improved substantially (we expect
800K events) allowing to study topics that have not been systematically analysed
and/or are plagued by sparse data.
A much complete description of MINERvA physics’ goals can be found on [2],
here we state some of the most relevant:
1
This MINERvA web page’s title describe exactly the objective of the experiment https://
minerva.fnal.gov/
• Precision measurement of the quasi-elastic neutrino-nucleus cross-section

including its Eν and q 2 dependence and study of the nucleon axial form
factor.
• Determinate the cross-section in the resonance-dominated region for both

neutral-current (NC) and charged-current (CC) interactions.
• Make precision measurements of coherent single-prion production in car-

bon, which is a significant background for next-generation of neutrino oscil-
lation experiments probing νµ → νe oscillation.
• Study of nuclear effects on sin2 θW measurements, and the NC/CC ratio for
different nuclear targets.
• Improve the measurements of the parton distribution functions with a ex-

pected sample of DIS events.
3.1.1 The NuMI Beam at Fermilab
The NuMI beam (Neutrinos at the Main Injector) is a beam of 120GeV protons
from the Main Injector that collide into a graphite producing secondary pions and
kaons, which are focused with two magnetic horns and directed into a 675m long
decay pipe where most of them decay (eq. 3.1) producing neutrinos and muons.
Muons are absorbed and monitored, through a total of 240 m of rock downstream.
Neutrinos finally reach to the MINERvA and MINOS experiment (fig. 3.2).
µ + → µ + + νµ , K + → µ + + νµ
µ− → µ− + ν¯µ , K − → µ− + ν¯µ
µ− → e− + νµ + ν¯e , K + → π 0 + e + + νe (3.1)
µ+ → e+ + νe + ν¯µ , K − → π 0 + e− + ν¯e
The NuMI beam provide around 1020 protons per target, by a hadron focusing
system it can produce energies of 1-3 GeV (low energy beam), 3-8 GeV or 8-
20GeV (more energy, higher pion energy and neutrino energy). Regarding time,
the beam is delivered in small package called spills with a duration of 10µs
long, every 1.6667 seconds. The change in the current of the horn allow the
experiments to have neutrinos and anti-neutrinos.
Figure 3.2: Diagram of NuMI and its different elements that produce the neutrino
beam used by MINERvA, MINOS and NOvA experiments. fig. by . Pavlovic.
3.1.2 MINERvA Detector
As mention in the sec 2.1, there are many channels for neutrinos to interact with
the nuclear targets depending of the energy and the type of charged final state
particles, whom enter into the tracking region and are finally detected. Muons or
pions have specific energy deposition footprints and it is the light produced in the
scintillator strips that are transformed into an electrical and digital signal ready to
be calibrated and reconstructed. In this section I will describe the most important
parts of the MINERvA detector.
Module Assemblies, Nuclear Targets, ECAL and HCAL regions
The module assemblies and nuclear targets are the core of the MINERvA de-
tector and where the interactions take place composed by 120 hexagonal-shape
modules suspended vertically and stacked along the beam direction. There are
four types of modules: tracking, electromagnetic calorimeter, hadronic calorime-
ter and passive nuclear targets.
The tracking modules consist of two scintillator planes, each composed of
triangular scintillator strips glued together, in two specific directions: U-axis and
V-axis specially defined for MINERvA detector and transversal to the axis parallel
to the beam named the the X-plane.
Respect to the x-y plane, the the U- and V-planes are rotated 60 degrees
clockwise and counterclockwise in order to avoid ambiguities with reconstructed
multiple-tracks hits associations. A squematic view of the detector can be seen
in the fig. 3.3
Figure 3.3: Side view of the complete detector showing the nuclear target, the
active region and the surrounding calorimeter regions. Image[2]
Nuclear targets
The most upstream part of the detector include five layers of nuclear targets
separated by four tracking modules that ensure a good vertex position resolution
for events originated in the nuclear targets. In the right side of the fig. 3.3 it can
be seen the area for the nuclear targets where neutrinos interact with.
There are 3 types of solid nuclear targets materials: lead (1014 kg, located in
targets 1,2,3,4), carbon (166 kg, located in target 3) and iron (976 kg, located in
target 1,2,3) that form the targets modules, the fourth one is an hexagonal-shape
pure lead plane, while the others contain mixed materials with different regions
of the detector. Also MINERvA has two liquid nuclear targets: water (500 kg
currently unfilled) and helium (250 kg of cryogenic liquid). In the fig. 3.4, it can
be seen the organization of the modules, and how they are mixed.
ECAL/HCAL regions:
The electromagnetic (ECAL) and hadron (HCAL) calorimeters region wrap
around the ouside the region trackers in order to contain the energy deposited
Figure 3.4: Active Scintillator Modules, and the 5 types of nuclear modules. Image
from A. Norrick [25]
and stop the electromagnetic (e.g. photons) and hadrons particles (e.g. pions)
correspondingly.
Ten tracking modules compose the ECAL region, this modules are 0.2 cm
thick lead collar wrapping around the entire scintillator plane and a 0.2 cm lead
sheet on the downstream end of the last plane in the module. While HCAL region
consists of 20 tracking modules are organized by one module of scintillator and
one 2.54 cm thick hexagonal steel plane in the inner detector region. In the fig.
3.3 (right), it can be seen where the ECAL and HCAL regions are.
One advantage of the calorimetric detectors over large Cerenkov detectors,
is that all particles all visible in principle, but the pay off is that is needed heavy
materials such as steel. This configuration produce as a results an outgoing lep-
ton plus pions and secondary particles, therefore the incoming neutrino energy
must be reconstructed from the muon and the energy of associated shower. This
technique has the downside that any unaccounted particle, either because a pion
was absorbed in the steel planes or within the iron nuclei themselves is expresed
as an error on the reconstructed energy scale.
Finally in the most upstream part of the detector, the Veto Wall is made of
two planes of scintillators which tag the events produced by the charged particles
outside the detector like rock muons.
Optical System
As mention before, the key concept in the MINERvA detector is the collection of
light pulses due to the interaction of particle products of the neutrinos interac-
tions with the nuclear targets. Light signals from more than 32000 scintillators
strips in the detector is converted into electrical signals which have amplitudes
proportional to energies deposited and carry timinig information.
The MINERvA Optical Systems is composed by three processes: the collec-
tion of the energy deposited in the tracking modules, the transformation of light
into an electrical signal, the readout electronics and one subsystem: the Data
Acquisition System.
The scintillators (fig.3.5) are strips of extruded plastic with a fixed triangu-
lar profile made of polystyrene pellets doped with 1% of 2,5-diphenyloxazole
(PPO) and 0.03% (by weight) 1,4-bis(5-phenyloxazol-2-yl) benzene (POPOP) and
a white reflective coating based on 15% T iO2 (by weight). Inside the strips, there
is a 1.2mm diameter wavelength shifting (WLS) fibers that collect the light and
transmit it into the DAQ system in one end. On the other end has been deposited
a 2500 Ȧ thick reflective coatting of 99.999% pure aluminium by sputtering [2,
p. 20].
Finally, the Fujikura-DDK optical conectors transmit the light from one end of
the fibers into the Hamamatsu Photonics H8804MOD-2 PMT boxes above the
detector (OD) with a spectral response of 300 - 650 nm and a peak wavelength
of 420 nm. The light output for a minimum ionizing particle (MIP) is redirected
into a low quantum efficiency photosensor within a 5 ns resolution that allow
to distinguishing the overlaping of NuMI beam spills, decay times or charged
mesons and time-of-flight measurements events. The readout process and the
DAQ is discused below.
Data Acquisition System (DAQ):
By DAQ systems we understand a set of interdependent components subsys-

(a) (Left) Cross-section profile of the scintillator strip. (Right) fibers

inside the strips.
(b) Fujikura-DDK optical (c) The complete PMT and its

conectors parts.
Figure 3.5: Images form [2]
tems that readout the data. MINERvA DAQ systems is formed by the readout
electronics and the Data Acquisition, which it main objective is to digitize the
electrical pulse signals through the FEB (Front Board Electronics), provide high
voltage to the PMTs and communicate with the computer interface readout sys-
tem (VME).
Some particularities have to be addressed. The systems must be able to dif-
ferentiate between NuMI spills of 10µs, work closely with the Accelerator Division
(AD) trigger systems and be robust enough to maintain a continuously readout
data during all the days of the year.
The former requirement is set since the MINERvA detector is trigger-less gate,
this means that we rely on AD signals for ”opening and closing the record of data”
in our detector and be more efficient in the managment of 32448 channels of data
(around 100kB/s) each 10 microseconds2 .
2
The main detector has 507 PMT, each one controls 64 channels (one channel per scintillator
strip).
(a) Front Board Electronics in MIN- (b) Time Profile of events in the NuMI beam
ERvA detector
Figure 3.6: Images by C.L. McGivern
Event formation and Calibrations
The event formation refers to the process that allow us to recognize the individual
interactions in time and energy. Since the detector is trigger-less the separation
of events has to be done offline, one for energy and one for the time that took the
event to happen.
The time-to-digital TDC data are first corrected for propagation delays to the
center of each scintillator strip. The fig. 3.7 shows a typical readout gate in the
MINERvA detector.
Figure 3.7: Typical time slice of . Screenshoot taken on 3/9/16
Time: an offline peak-finding algorithm create ”time slices” (the peaks with
different colours seen in fig. 3.7. If during 80ns after the hit, which fired the
discriminator, the energy of hit is 2/3 of the signal over a plane for a normally-
incident minimum ionizing particle, then the time slice continue to be recorded
until the energy condition is not longer met. A single neutrino interactions are
usually contained in a single time slice.
Energy: The raw analog-to-digital converter ADC data must be calibrated to
provide an estimate of the energy deposited in each scintillator strip, but consider-
ing the correction needed since there are four effects must be taken into account.
(A) Attenuation of photons while they travel along the wavelengtth shifting fiber to
the end of the strip, (b) the light signal is attenuated in the clear optical fibers, (c)
PMT gains in the amplification of the photoelectrons and (d) the application of the
ADC conversion function in the digitalization by the FEB of the electrical signal.
The estimation of the energy with all the correction can be found in more detail in
[2, p. 35].
Calibrations: Since we are measuring time and energy, we need a reference
point from where we set the zero level, but because we are working with a dy-
namical system, the calibration process is made in order to continuously correct
the true zero value for the 32000 channels.
There are two types of calibrations: in situ (dynamic measurement), and ex
situ (static measurements). The ex situ calibrations were made in order to char-
acterize the tracking region, for example measuring the attenuation length in the
shifting fiber and apply the rightful correction for obtain the true energy.
An in situ calibration is performed during the readout process, in order to
dynamically set the zero value or the background voltage of the PMT without
interactions.
3.2 Test Beam Detector
The MINERvA’s Test Beam experiment is a scale-down detector replica of solid

scintillator tracking and sampling calorimeter regions used in the Fermilab Test
Beam Facility (FTBF) which recieve particles of know momentum and type. These
particles are produced by a beam of 120GeV protons that hit an aluminum target,
creating secondary beam of particles (e± , π ± , p± , K ± ). A set of focusing mag-
nets select the momentum of the different particles that arrives to our detector
with energies from 1GeV to 16 GeV. In the fig. 3.8 it is shown the location of the
facilities.
(a) Map of the location of the FTBF at Fermilab.
(b) Buildings of the FTBF with two versatile beamlines (MTest and MCenter) in which the
experiments like MINERvA can run a full Detector R&D experiment.
Figure 3.8: Images of the FTBT.

3.2.1 Why is necessary to have a Test Beam program
There is a continuous requirement of improvement the simulations models of the

calorimetric response of single state particles in the MINERvA detector. And the
validation of theses models has been conducted in a test beam program at the
Fermilab Test Beam Facility in 2010 for low energies and 2014-2015 for medium
energy[3].
A tertitary test beam with hadron momenta between 0.4 and 2.0 GeV/c was
used in the low energy in order to study the response of a small-scale down
MINERvA detector with the final particle state from neutrino interactions with the
nuclear targets. An auxiliary detection system has been placed in order to identify
the momenta, the direction and the identity of the incoming particles. The data
acquired was compared with a Monte Carlo simulation of the testbeam geometry
and by using the same software and calibration infrastructure of the MINERvA
detector the models of calorimetric response are improved. In the fig. 3.9 we can
see that by increasing the of positive pions, the Test Beam 1 detector reduces its
response below to 60%, a similar insight is shown on the right side for negative
pions.
Figure 3.9: Calorimetric response for positive (left) and negative (right) pions for
low energies in the Test Beam Program 1 for low energies.
In 2014-2015, a second stage of the MINERvA Test Beam program has been
conducted in order to study the models and the calorimetric response in the
medium energy using the seconday beam at MTest. My participation at MIN-
ERvA has been concentrated in a small analysis regarding the time structure of
the income beam at the FTBF. The final results of the Test Beam for medium
energy are still been analyzed.
3.2.2 FTBF Beam production
From the Proton Source, 750keV H − ions are extracted from the LINAC (A in fig.
3.8) that accelerates them up to 400MeV. As the ions are injected into the Booster
(B), the electrons are removed leaving protons to circulate into the Booster and
be delivered to MicroBooNE (2.3.2) or continue through the Recycler and Main
Injector (C) where the protons are accelerated to 120GeV (C) with frequency
of 53MHz. At this point (D), the beam can be delivered to the NuMI (E) and
experiments like MINERvA (F) or NOvA, or to the FTBT (G).
As can be seen from the fig. 3.2.2 (up), the 120GeV protons hit an aluminum
target (G), producing e± , π ± , p± and K ± that are used in the MTest [I] where the
experiments place their scale-down detectors, and in particular the MINERvA’s
Test Beam detector 2 (from now on TB2).
The beam at the FTBF, has a inner structure regarding the time as a variable
composed by three scales that allow us to describe the time profile of the beam4 :
Bucket: this time correspond to the RF rate in which one particle (ideally), is
extracted and accelerate in the Radio Frequency Cavity. The duration of a bucket
is the 19ns. A collection of 84 buckets (1.6µs) form a batch.
Batches: lenght in time of one Main Injector cycle that is form by 7 batches
(11.2µs).
Spill: the time of resonant extraction of the beam from the Main Injector over
375000 MI Cycles, to create a 4.2 seconds spill ($21 event) on every rotation the
beam make around the machine.
Chapter 4 is dedicated to explain time beam structure, and appendix C review

radiofrequency cavity’ s theory in more detail.
4
In the case of NuMI, the beam spills are around 10µs ong every 1.667 seconds DAQ triggers
off of $A9s signals
(a) Creation of the secondary beam in the FTBF. Image made by A. Norrick.
(b) Inside the FTBF
Figure 3.10: Path that follows the beam inside the FTBF before enters into the
MINERvA’sTest Beam Detector 23 .
3.2.3 Test Beam detector’s auxiliary systems
The MINERvA’s Test Beam 2 detector have two calorimetric regions (ECAL and
HCAL) which resembles the ones in the main detector, in order to reproduce the
conditions of the final states of neutrino interactions in a controlled enviorment

and so forth improve the models of detections of p± , e± , π ± and µ± as mentioned
in the sec. 3.2.1 within the range of 2GeV through 16GeV both polarities.
During 2015, the TB2 detector has been receiving electrons and hadrons, the
premliminary numbers of amount of particles can be looked in the fig. 3.11 and
3.12.
(a) ECAL/HCAL Negative Beam (b) ECAL/HCAL Positive Beam
Figure 3.11: Energies and Polarities taken in the Run 1. Tables made by R. Fine.
(a) Tracker/superHCAL Negative Beam (b) Tracker/superHCAL Positive Beam
Figure 3.12: Energies and Polarities taken in the Run 2 and 3. Tables made by R.
Fine.
The data that has been taken is divided according a two configurations of
the calorimetric regions (ECAL/HCAL and Tracker/SuperHCAL ) of the TB2 de-
tector and the chronological time in which it was recorded (Run 1, Run 2 and
Run 3). The CAL/HCAL configuration is composed by 20 planes of scintilla-
tor/PB and 21 planes of Scintillator/Steel, while the Tracker/SuperHCAL configu-
ration: 20 planes of Scintillator, 4 planes of Steel/Scintillator, 11 planes of Double
Steel/Scintillator and 6 planes of Steel/Scintillator. The fig. 3.13 shows this.
The TB2’s auxiliary detection system has been choosen in order to recog-
nize single particles of known type and momentum in a calibrated detector, the
components can be seen in fig. 3.14: four Wire Chambers (L), Veto System (M)
the Time of Flight System (ToF J and N) and the Cerenkov detector (K). We will
describe the different parts of the systems.
Figure 3.13: The two configuration for the calorimetric regions in the TB2. Image
made by A. Bercelli.
(a) Tracking logic of the TB2 auxiliary detection system. Image made by R. Fine,
the annotations are ours.
(b) Particle ID logic. Image by Anne Norrick.
Figure 3.14: (Left) Schematic view of the Test Beam Detector and the Auxiliary
detection System. (Right) Components of the Test Beam and its functions.
The Time Digital Converter TDC/CAMAC Lecroy 3377
The TDC (Time to Digital Converter) CAMAC Lecroy 3377 (CAMAC 3377 from
now on) is a machine that provide a digital representation of time when an event
has ocurred. The CAMAC 3377 provides us with a high-resolution time measure-
ments for low measurements dead times. It has 32-channel, each individual input
channel has a LIFO (last in, first out, a time of processing the data) type buffer
attached to it such up to 16 hits can be recorded on the channel with respect to
the common hit.
The CAMAC 3377 is the interface between the auxiliary system (the Veto Wall,
ToF, Wire Chambers) with the Data Acquisition System readout information from
the ToF and Veto systems, control trigger sent to the MWPC and identify triggers
in spill and out of spill.
The multihit capabilities allow us to record multiple events in the same win-
dows of time (that we set). Its pulse width measurements is up 16bit dynamic
range with 500psec of last significat beat, which mean that the minimum window
of time that we can reach with the CAMAC 3377 is 500ps. The time is meaured
with respect to a common reference time mark or ”COMMON HIT”, that can occur
before or after the infividiual time signal.
Some other characteristics:
• 32-channel multihit TDC: we can measure up to 32 channels in the same

window of time
• Each channel can measure up to 16 measurements on each channel (512

events per window of time)
• 500ps digitizing resolution (LSB): this is the maximum resolution that we

can achive
• 32 µs of time full scale: we can construct our windows of measurements in

multiples of this value
• 8 trigger outputs programmable

Table 3.1: Modes of operation of the CAMAC 3377 TDC. The ? means that there
is no information
Mode Off set resolution Time range Full scale time range
Single word 500ps - 4 ns 255 ns - 4µs 0 - 32µs
Double word 500ps 8ns - 32.7µs 32.7 µs
Mode Off set resolution Steps Full scale time range
Standard no info 8ns 8 ns - 32 µs
Common Start mode no info 50 ns up to 32 µs
Common stop mode no info no info no info
• The number of edges measurements recorded per channel is 1 to 16
• Multi event buffer allows up 31 small events and 4 full events can be recorded
before readout
• The readout can occur in background, while the fron end is recording data
hits.
• The readout is by event
• Modes of operation shown in the table 3.1
The double word format preserves the full 16 bit data for wide dynamic range.
At the end of acquisition the data is unloaded from the MTD133s and stored in a
multievent FIFO buffer. Dead time is 1.8µs plus 100 ns per recorded hit (200 ns
per hit when in double word mode).
Veto System
The Veto System is a set of 12 scintillator paddels that surrounds the central re-
gion and looks for particles entering the detector outside of the direct beamline. A
300 ns resolution has been set up since this is the value the maximum resolution
that the MINERvA detector have to recognize two different events.
As will be explain in the sec. 4.3, in theory the AD deliver zero or one particles
of know energy per MI bucket, but in practice this is no the case. Sometimes send
more of one particle in single MI bucket, or more than one particle in adjacent
buckets/batchs. Since the detector needs to know that the input is one particle of
know energy, the Veto System allows to tag all other events.
(a) (b) Main Injector Spill seen by the CAMAC 3377 during a normal data
taking.
Figure 3.15: (a) TDC-CAMAC Lecroy 3377 used in NuTeV and CDF.
Figure 3.16: Veto system and its six paddles. Photo by A. Norrick
The Wire Chambers and Veto System are responsible for detection of single
particle events and discard multi-particle events. The Wire Tracker 4 (MWPC4)
centered on the central value of the beam position in X and in Y allow us to
know if a particle is on the axis beam, but for other angles the mentioned array of
scintillators counters, all the space for approximately 1/2m around the beam axis
send a signal if some particle hits them.
Time of Flight (ToF)
The Test Beam experiment is able to measure the time that it takes for a particle
of known momentum to travel between two known locations. From this, it can
be calculate the mass of the particle and consequently its identity. The system
is formed by two stations with PMTs that record the time that a particle passed
through them. The stations are 104.5 m apart, the dowstream station have two
paddles while the upstream; four PMTs as shown in the fig. 3.17.
If a particle is sending through the detector, the ToF system will receive the
trigger signal which opens the gate for all the six different PMTs independently in
both stations. The channels will continue counting until the stop signal is received.
And, the security condition is placed in order to stop recording other signal.
Figure 3.17: Time of Flight Upstream and Downstream detectors. Photos by A.

Norrick
The rise time of the two downstream PMTs is 3.0ns, with a jitter of about
0.4ns, while the upstream PMTs has a rising time of 1.3ns and a dead time of
about 0.3ns5 [5].
Cerenkov detector
The Cerenkov dectectors (2) are gas chambers with PMTs attached to them,
that when a particle emits a characteristic cone of light. Particles moving slower
than certain thresholds speed are invisible in water Cerenkov, but fast charged
particles may create electromagnetic wave while traveling along a medium, this is
called Cerenkov radiation if the particle velocity is greater than the velocity of the
light in that medium. The radiation generated is spread in the shape of a cone,
along the direction of the particle.
5
In sec. C.3 we outline very briefly the physics of the ToF.
Figure 3.18: A diagram of the Cerenkov detector used at MTest (Fermilab). Copy-
right Fermilab.
The following equations 3.3, show the relationship between the angle in the
cone with the velocity of the particle and the index refraction. While in the other
equations it can be seen that by changing the pressure and assuming an isother-
mal behavior, a simple spring model of electrons in a diaelectric medium describe
the phenomena, a threshold pressure and a relation between pressure and the
refraction index is found.
1
cos(θc ) = (3.2)
βn
q 1 −1
2
P1 ρ1 n1 − 1 1− m2
E
= = PT = (3.3)
P2 ρ2 n2 − 1 δ
where δ = n1atm − 1 = 0.000297 and PT is in atm.
The Cerenkov detectors at MTest can be used as threshold counters (particle
ID) in order to tag the electrons in the beam. There are two Cerenkov counters at
MTest, one upstream is ’80 long and one downstream, ’50 long. They have two
PMT attached that collect the light form the cone and amplifies the signal, while
subtracting the noise with help of other equipment.
The highest momentum at which a particle can be identified is determined
by the velocity resolution of the counter combined with the characteristics of the
beam. While the lowest momentum that can be tagged for a given particle is set
by the particular gas used and the pressure. In the table 3.2 it can be seen the
value on GeV for two different gases and particles.
Table 3.2: Lowest momenta value for detection in Cerenkov detectors at MTest
(Fermilab)
electron muon pion kaon proton
Nitrogen 0.02 4.0 5.0 1.8 35
C4F8O 0.01 1.8 2.4 8.0 15
Figure 3.19: Spectra of de Cerenkov detectors at MTest for the MINERvA Experi-
ment. Work done by M- Ramirez
Wire Tracking Chambers
When a charged particle passes through the material (the two planes at 90 de-
gree angle) a signal is generated with the xy position for each hit. The ability of
track in space the particles and with the help of 4 wire chambers in the detection
system, allow us to track particles as they go into the detector.
3.2.4 How an event is recorded: trigger logic and DAQ for

MINERvA Test Beam 2
In all the measurements in physics, time is an important parameter that allow not
only to study the dynamic relationships but also structure the way the data is tak-
ing. In this last section I will describe how the detector (and all their subsytems)
are triggered, this mean the physical signal that tells the detector this is an event
that I am interested in recorder for further analysis. In specific, the FEBs are the
ones that open and close the data readout and increase the voltage in the PMT
and integrate the charge into a digital signal.
Triggers in the Main Detector
The Main Detector is triggers-less, this means that we rely on other systems for
opening and closing the process of taking data. In our case is the Accelerator Di-
vision through the $39 signal that start this. Every 1.67s (or 1.33s) the MINERvA
detector receives 10µs neutrinos spill.
From the beginning of the signal, the time that takes for setting the correct
high voltage into the PMT is 20µs, and it ends just before the beginning of the spill
mentioned before. The FEB integration is closed after 5.5 µs after the neutrino
pulse and from this point the Data Acquisition begins.
Figure 3.20: Underground trigger timing with the MINERvA Main Detector.
When a burst of neutrinos reach the MINERvA Main Detector, the FEB are
open hoping to read an interesting event. This does not happen in the Test Beam
detector.
FEB Gates at the Test Beam
In the case of the Test Beam, the timing trigger is different since the time structure
of the beam is different from NuMI and we, in theory, receive individual particles
instead of a neutrino pulse. Here, the FEB gate is open as often as possible and
hope that a particle hits the detector when the FEB gate is open.
Each minute, we receive a slow-spill of 4.2s, having also a signal from Accel-
erator Division telling us that beam is coming to the test beam detector. But, the
trigger (the signal that difference data from no-data) is constructed with an Aux-
iliary System. By using the electronics from the Main Detector, the HV settling
time and the integration gate is the same, the differences are that FEB integration
gate have an overlapping with the HV settling time of 1µs, an 9µs of window time
in which if a particle is detected the Data Acquisition will start6
Figure 3.21: Underground Trigger timing within the Test Beam Detector.
Since, the Test Beam detector is always ”open”, some no desire events will
be detected and needed to be tagged in order to remove it from the final data.
This is done through conditions set up with the auxiliary system. There are two
types of triggers: beam and cosmic trigger. The last one because, from time to
time, cosmic particles will reach the detector. This events are used for internal
calibrations. The readout is done by the CAMAC and correspond to the variable
out-spill in the ROOT file.
Regarding the beam trigger, its main objective is to set the signal that a parti-
cle during the main injector spill is aimed at the detector. Constructed by a logic
between the auxiliary detection system (Veto, Cerenkov and CRIM).
If we have a cosmic particle out of spill or a particle from the beam, with the
confirmation that it is not a electron (in this case the trigger output is restricted
6
Remember that in the Main Detector, a minimum energy during a specific window of time has
to be reach in order to consider that event as a neutrino interaction event.
(a) Events and logic that do fire the Test Beam DAQ to begin the readout
of the data. Image by G. Savage.
(b) Events and logic that do not fire the Test Beam DAQ to begin the read-
out of the data. Image by G. Savage.
Figure 3.22: Logic of the triggers in the Auxiliary Test Beam detectors.
by the Cerenkov and spill signal) and that did not hit the Veto System, then that
particle will tagged as data. Therefore the Veto inhibit the DAQ to readout triggers
for these conditions: Trigger (allow CAMAC readout to complete), the begin and
end of the spill. The CAMAC DAQ is responsible for mapping of out-spill/in-spill
triggers in MINERvA DAQ.
Chapter 4
Time structure of the FBTF Beam
As mentioned in sec.3.2.1, the MINERνA experiment is composed of two detec-

tors places in different parts at Fermilab, with different physics’s goals1 . We have
describe the production of the NuMI beam that is used for neutrino cross-section
scattering in the main detector, and in the sec. 3.2 we did the same for the FTBF
Beam, that is used in the MINERνA experiment.
This chapter is meant to give an overview of the concepts involve in the pro-
duction of a beam of particles that is use in the experiments at Fermilab, and spe-
cially at MINERνA experiment. We start with general concepts that are needed
to understand how the beam is produced (sec. 4.1), then we describe the Fer-
milab’s Accelerator Complex (sec. 4.1.1) which produce the beams used in the
experiments. One of the most important elements in modern accelerators is the
Radio Frequency Cavities (RF) which accelerates the particles; an introduction
modes in the RF, the Q factor of the cavity and the Phase Focusing hability of the
RF systems are described (sec. 4.2), in order to explain in the last section (sec.
4.3) why the beam has a defined structure2 .
The approach followed in this thesis regarding the details of the production
and delivery of the beam is focus on the periods of cycles of production of the
beam that we receive at the FTBF. Topics of physics of beamlines, monitoring,
control, production and instrumentation are away from the scope of this work.
1
The principal detector is placed at the NuMI Hall 3.1.1 while a small replica is placed at the
Fermilab Test Beam Facilities. More information can be found in the sec.3.2
2
As a remainder the aim of this thesis is to show that the MINERνA’s Test Beam Experiment
is able to resolve two of three scales in the beam structure
54
CHAPTER 4. TIME STRUCTURE OF THE FBTF BEAM 55
4.1 General Concepts in Accelerator Physics
Beam of particles
A beam of particles are a concentrated group of charged particles, that are ac-
celerated by increasing their kinetic energy in particle accelerators. While beam
intensity is defined as the number of particles in the beam that we can measure
as the number of particles. Usually the beam direction is defined along two di-
rection in order to describe the particle motion: the longitudinal dimension (the
direction in which the beam travels) and the transverse dimension that it is form
by the horizontal and vertical axes and form the transverse plane.
An accelerator is a machine that accelerates particles over a trajectory, and
are use with the help of other machines in order to deliver a specific type of beam.
Roughly speaking there are two types of accelerator: fixed-target accelerators
and colliders. The former, produce a beam that is smashed into a fixed targed
which produce secondary particles that are used by the experiments. In the later,
two beams are guided and then collided, the particles that results from this are
studied.
One of the most well know collider laboratory is the CERN’s Large Hadron
Collider (LHC), while the most powerful neutrino beam in the world is produce by
NuMI’s Fermilab. As mentioned in sec. 3.1.1, the NuMI beam is produced when
120GeV protons collide into a carbon target.
The two principal elements of an accelerator are: radiofrequency cavities,
which produce the acceleration of the particles, and magnets. Whom change
the direction of the beam with the help of the force exerted into a charge particle
within an electric and magnetic field. Magnets, bend the direction of the particles’
trajectory, dipoles change the trajectory of an entire beam while quadrupoles
magnets, focus or defocus the beam. Usually, the magnets are arrange in a
lattice of different types.
Momentum and energy of particles
v
Since the particles travels with velocities near the light’s velocity, β = c
, the
energy total energy is ET otal = γmc2 . A proton extracted from the Main Injector
at 120.00 GeV, have a energy at rest (Erest = mc2 ) of 938.26 MeV so forth it will
have a velocity of 99.997%c.
r
120.00GeV 120.00GeV 2
γ= β= 1−( ) = 0.99997
938.26M eV 938.26M eV (4.1)
Dynamics of accelerators
The only way to interact with electric charges is through the electromagnetic
fields. If a particle has a velocity ~v the force exerted on a particle of charge q
is:
F~ = q(E
~ + ~v × B)
~ (4.2)
From this equation, we can see that the force due to a static magnetic field
is perpedicular to the particle’s velocity. By the Work-Energy Theorem, this force
won’t increse the kinetic energy of the particle.
Lets assume that a particle is moving along a curve through an electromag-
~ → Eθ and B
netic region with where the fields have single components E ~ → Bz .
The Newton-Lorentz force in a curvilinear coordinate system reads:
d~p ~ + e~v × B
~
= eE (4.3)
dt
d(mvθ ) v2
· ~uθ − m θ · ~ur = eEθ · ~uθ + evθ Bz · ~ur , (4.4)
dt ρ
where ρ is the local radius of the trajectory. It can be see electric field provides
energy and momentum to the particle where magnetic fields bends the particle
trajectory.
pθ
= eEθ
dt (4.5)
pθ /e = Bz ρ
4.1.1 Fermilab’s accelerator complex
In the fig. 4.1 it can be seen the machines that compose the Fermilab Accelerator
Complex composed by: the Ion Source, the LINAC (A), the Booster (B), the Main
Injector (C) and the different outputs for low-energy neutrino experiment, muon
experiment, high-energy neutrino experiments and the Test Beam Facility.
750keV H-minus ions are extracted from the source into the Linac. The Linac
accelerates the ions to 400MeV, and then extracts them to the Booster Accelera-
tor. As the ions are injected into the Booster, the electrons are stripped off leaving
400MeV protons to circulate in the Booster.
The LINAC is formed by RF cavities placed in-line with one another in order to
provide a large amount of energy gain per unit length. The beam only pass onces
because is linear and since the beam is composed by similar particles, they tend
to electrostaticaly repel one another, which generate a spread in the beam size
and has to be corrected with the help of quadrupoles.
The Booster is a 1500-foot-circumference rings that accelerates the beam up
to the energy of 8GeV, which provides a low-energy neutrino beam for the Main-
Injector and to low-energy neutrino experiments like MiniBooNE. The Booster
captures the protons into 84 bunches (1 batch) and accelerates them to 8GeV.
Each of these bunches is 19nsec long. Typically, 8 - 30 of these bunches are
extracted to the Main Injector (MI) for Test Beam operation (a process known as
Partial-Batching) At the injection total energy of E = 8.938 GeV, the Main Injector
has a circumference in time of 11.13 µs, which is exactly 7 booster batches long.
The Main Injector, situated directly beneath the Recycler in the same tunnel,
ramps up to proton beam from the Recycler from 8GeV to 120 GeV. Being a
circular accelerator know as ”synchrotron”, that is compose by a magnet system
and a RF system. Both are ”synchronized” as the kinetic energy of the beam
increases. The Recycler, is a stage area for proton beams after exits the Booster.
(a)
(b) Fermilab Accelerator Complex and th principal outputs from the Main
Injector, to the different experiments.
Figure 4.1: Fermilab Accelerator Complex.
Here is where the beam is combined into batches of protons to form a more
intense beam. Once it is done the proton enters the Main Injector, on top of the
Recycler.
The Main Injector accelerates the beam to 120GeV at a frequency of 53 MHz,
at which point a process called Resonance Extraction is started and a fraction of
the beam is resonantly extracted in a slow spill for each Main Injector rotation.
The machine cycle is the cycle, in which a defined sequence of task are
performed in regular intervals, and it is shared by all the machines that compose
the accelerator complex. The entire machine cycles at a fixed rate of 15MHz, this
means that all of the equipment performs a given task fifteen times a second.
This is called a ”beam cycle”[12].
4.2 Radio Frequency Systems (RF Systems)
A RF cavity is a electromagnetically resonant structure that generates a strong

longitudinal electric field that accelerates beam and does not affect the orientation
of the longitudinal or transverse dimensions. Usually working around of 3 kHz
to 300 GHz, modern particle accelerators use them in closed geometries that
produce standing waves and increase the kinetic energy as the beam passes3 .
This section is devoted to explain the basic operation of the RF cavities, since
them play an important role in the inner structure respect of time of the beam. In
the appendix C we review the electromagnetic theory of the RF systems, while in
sec. 4.3 we connect how the RF systems impact in the time structure of a beam
in synchrotrons.
4.2.1 RF Cavities
EM in vacuum or matter accept monochromatic plane waves that propagate along

the space with a well define velocity. In both cases, when the surface is reached,
the EM wave will have a reflected, transmitted and incident wave.
EMW in conductors consider the existence of free charges and currents, that
we don’t control ρf and J = σE will impose the following boundaries conditions:
3
If we where using DC voltage, electric breakdown at the HV terminal is one of the main
problems to overcome when a particle is accelerated, instead for modern accelerator it used
AC generator of RF waves. Only for the Main Injector, the required voltage is 120E9 while the
dielectric strength of the air is 3E6. In order to avoid the arcing between electrodes the separation
between them would need to be of 40 kilometers in a ”static” the Main Injector. [12]
k k
Ē1 = Ē2 B̄1⊥ = B̄2⊥
1 k 1 k (4.6)
1 Ē1⊥ − 2 Ē2⊥ = σf B̄1 − B̄2 = K̄f × n̂
µ1 µ1
Which in the case of a perfect conductor the incident wave will be totally re-
flected with a −180 phase shift.
Figure 4.2: A niobium-based 1.3 GHz nine-cell superconducting radio frequency

to be used at the main LINAC of the International Linear Collider. Photo from
FNAL.
The generation and transmission of electromagnetic radiation involves metal-

lic structures with dimensions comparable to wavelengths (meters) that we are
working with. Hollow metallic cylinders that produce the propagation or excitation
of electromagnetic waves are called wave guide. From the Maxwell’s equations,
and considering a axial symmetry a general solution can be found.
ω2

E
[∇2t 2
+ (µ 2 − k )] =0 (4.7)
c B
The explict solution is composed from a tranversal electric and magnetic fields
with the boundary conditions imposed to the normal component to the surface of
the electric field and the parallel component for the magnetic field since the wave
guide surface is a perfect conductor.
1 ∂Bz ω
Bt = 2 [∇t ( ) + iµ k̂ × ∇t Ez ] (4.8)
(µ ωc2 − k2) ∂z c
1 ∂Ez ω
Et = ω2
[∇t ( ) − iµ k̂ × ∇t Bz ] (4.9)
(µ − k 2 )
c2
∂z c
As can be seen, in order to fully calculate the fields, it is only needed the Ez
and/or Bz . The two boundary conditions can not generally be satisfied simulta-
neously, therefore we have two distinct categories of EMW that can exist in the
wave guide: a tranverse magnetic mode (TM) and a transverse electric mode
(TE).
Transverse Magnetic (TM)
Bz = 0, everywhere. Boundary condition: Ez |S = 0
Transverse Electric (TE) (4.10)

∂Bz
Ez = 0, everywhere. Boundary condition: |S = 0
∂n
The logic choice in order to accelerate charge particles is to use the TM mode,
since the Bz is zero everywhere, avoiding the change in the trajectory of the
particle (eq. 4.5).
Furthermore, the boundary conditions constrains to a spectrum values that
γλ can take. Each λ is called modes of the guide, and the frequency of the
electromagnetic planes (ω) is determined according to the values:
ω2
kλ2 = µ( 2 − γλ2 ), (4.11)
c
ωλ is defined as the cutoff frequency, since :
γλ
ωλ = c √ (4.12)
µ
and the wave number can be written as:
1√
q
kλ = µ ω 2 − ωλ2 (4.13)
c
it can be seen that, for ω > ωλ the wave number is real and the wave propa-
gates through the wave guide. When it is not positive, kλ is imaginary, and it is
attenuated while it propagates, which is not the desired case for RF Cavities.
At this point we know that RF wave guides generate an electric field which
in the axis of the guide is parallel to the axis, but a cavity has ends, so forth
an additional condition is impossed to eq. C.16. The cavity’s walls are taken to
have infinite conductivity, while the cavity is filled with a lossless diaelectric with
constants µ, .
Figure 4.3: A pillbox cavity. The lower mode frequency does not depend from the
height of the cavity.
The reflections on the ends of the cavity add additional boundary conditions,
which means that the waves will be reflected. The general solution for TM waves
are:
TM waves
pπ pπz
Et = − 2
sin( )∇t Ez
dγ d
iµ pπz (4.14)
Bt = − 2 cos( )k̂ × ∇t Ez
cγ d
For a pillbox (a cavity cylinder) the TM mode the transverses equation have
solution that includes Bessel functions and an angular dependence in φ. For
ψ = Ez and with the boundary conditions Ez = 0 at ρ = R, the solution for the
lowest resonance frequency in TM mode (m=0,n=1,p=0):
2.405ρ −iωt
Ez = E0 J0 ( )e
R
2.405ρ −iωt
Et = E0 J0 ( )e (4.15)
R
√ 2.405 −iωt
Bφ = −i µE0 J1 ( )e
R
The resonance frequencies depend on three indexes one from the periodicity
of the cavity and two from Bessel’s solutions.
r
c x2mn p2 π 2
ωmnp =√ + 2 (4.16)
µ R2 d
As can be seen in the the eq. 5.15 the electric field points longitudinally, the
magnetic field has a minimal transverse effect on the beam, while the energy is
stored and moves back-and-forth between the electric and magnetic fields with a
phase difference of 90 degrees.
4.2.2 Power Losses in Cavity
Resonant cavities have definite field configuration for each resonance discrete
frequency of oscillation. Fields will not built up unless the exciting frequency
matches the resonance frequency, but in reality there is a narrow band of frequen-
cies around the eigenfrequencies where excitation occurs. The quality factor or
”Q” measure the energy efficiency of an oscillator or the sharpness of response
of the cavity to external excitation:
Stored energy
Q = ω0 (4.17)
Power loss
Assuming ohmic losses, the stored energy decay exponentially according to

e−ω0 t/Q , where for larger values of Q, the decay is slower than small values of
Q (fig. C.3). Accepting that there is no single frequency but a superposition of
frequency around ω = ω0 , there will be a range of frequencies that will gener-
ate standing TM waves in the cavity. For this range of frequencies, the energy
distribution in the cavity can be calculated:
1
|E(ω)|2 ∝ (4.18)
(ω − ω0 )2 + (ω0 /2Q)2
which has a Lorentz line shape shown in the fig. C.3 with a full width at half-
maximun equal to ω0 /Q. The energy of oscillation in the cavity will follow the
resonant curve in the neighborhood of the particular resonant frequency. ∆ω is
the frequency separation between half-power points, so Q can be defined also
as:
Figure 4.4: The resonance curve’s full width is equal to the central frequency ω0
dived by Q.
ω0
Q= (4.19)
∆ω
This definition makes explicit that Q is related to the frequency width of the
cavity response. Since RF cavities with high-Q have a narrow frequency re-
sponse (a lower frequency width). This feature of RF cavities are useful for linear
accelerators, where the cavity frequency does not change. The Fermilab’s ac-
celerator complex (sec. 4.1.1) is composed by two synchrotrons (the Booster
and the Main Injector) that require the RF increase with the beam energy that is
achieve by attaching small coaxial RF transmission lines to the RF cavities that
are loaded with ferromagnetic material. This ferrite tunner change the inductance
of the entire system, altering the resonant frequency of the RF cavity while it is
needed in the synchrotron.
4.3 Time structure of the Beam
4.3.1 Synchronicity condition
Since the RF cavities are connected with other machines along the beamline, the
RF oscillations must match in time with the arrival of beam in the cavity, and take
advantage of the right force direction, during half of the oscillation. The particle
arrival time with respect to the RF cycle is know as the RF phase. By doing this,
the particle will always see an accelerating voltage at each RF gap.
To achieve this, the distance between cavities in a linear accelerator and the
RF frequency have to be chosen to prevent the beam to arrive in the cavities on
the other half of the oscillation. For this to occur, the following relationship must
be hold:
nν
ω0 L = , (4.20)
2
where ω0 is the RF frequency, the distance between cavities L, the particle
velocity ν, and interger n. This equation is know as the ”synchronicity condi-
tion”.
In the case of synchrotrons like the Main Injector, the RF cavities are placed
along the circumference of the machine and the frequency of revolution along it,
must be an interger of the RF frequency. The interger multiple of the revolution
frequency, h is call the ”harmonic number”.
ω0RF = ωREV (4.21)
Considering a particle with speed v = βc circulating along the machine with

period of revolution:
2πR βc
Trev = ωREV = (4.22)
βc 2πR
We can see that it depends on the radius and the velocity that at the end we
want to reach β, there is a maximum numbers of ”spots” that can be accelerated
on a synchrotron4 .The segments of the circumference centered on these points
are called buckets.
4
In the case of the LHC, there are approximate 35640 buckets but not all of them all filled with
particles with only 2808 of them occupied.
4.3.2 synchrotron Oscillation, Buckets and Bunchs
A particle that is exactly synchronised with the RF frequency is called synchronous

particle, but there are always slightly deviations in the particles momentum in the
beam and particles with different velocities will not met the synchronicity condition
4.20.
That is why the beam will have a non-zero energy and phase spread, because
each particle will arrive at a slightly different time, feeling a different electric field
strength.
Lets consider the case of three particles arriving on different times into a RF
cavity. The early-arriving particle (A) will feel a weaker electric field than syn-
chronous (B) particle, moving it toward the synchronous phase. For later-arriving
particles (C) will see a higher electric field that help them to adjust its velocity into
phase synchronous particles speeding up.
All particles will oscillate longitudinally around the synchronous particles un-
der the influence of the RF electric field, this longitudinally-focusing process is
called phase focusing. However there are limits to the phase focusing ability of
the RF, a particle outside the range of synchronous phase will not be pushed to
maintain stable oscillations. A bucket is defined as the stable RF space that can
be phase focused. For all the particles in the bucket, the beam has a net accel-
eration, phase focusing adjust the beam momentum for stability: this longitunial
motion is called ”synchrotron oscillation”.
In the case for synchrotrons the number of RF buckets in a machine is limited,
because the number of RF buckets will depend of the time it takes a particle to
make one orbit. The phase focusing process causes beam to be collected into
discrete packets known as bunches. Since the bucket area is the only stable
place for beam to exist, it is often said that the bunches fill the buckets.
For high energy particles in synchrotrons, they will have longer orbits and a
lower revolution frequencies (delaying its arrival at the RF cavity). Reciprocally,
low energy particles will have a shorter orbit, reaching the RF cavity sooner 5 .
Particles in the same time than synchronous particles but with higher energy will
5
Since the angular frequency is defined as ω = v/r, where r is the radius and v the velocity of
the particle.
see a decelerating electric field, reducing its energy and velocity until is enough
to surpass the correct revolution frequency, this particles will be accelerated by
the RF cavity reaching the same situation than the first turn. As can be seen,
synchrotron oscillation is inherent to the RF cavity. A beam is called unbuched
beam (DC beam) if it does not comes in bunches.
Figure 4.5: Graphic description of a Bucket and a Bunch, the RF voltage that the
particles see while they are in a bunch and how the buckets arrange themself on
a synchrotron’s ring.
4.3.3 Time Structure of the Beam at MCenter
The beam at the FTBF, has a inner structure regarding the time as a variable
composed by three scales that allow us to describe the time profile of the beam
Table 4.1: Time structure of the Beam according to Accelerator Division.
Scale Composition Related with Duration
RF Bucket ”one” particle RF’s frequency 19ns (52.8MHz)

Batch 84 RF buckets Booster length 1.6µs
MI cycle 7 Batches 1 cycle of the Main Injector 11.2µs
Spill duration 375000 MI cycles Resonant extraction 4.2s
We will explain in detail why it is structure in this way, and the role of the
different machines mentioned in the sec. 4.1.1 in the production, composition
and the duration of this scales.
It is necessary to mention that, the protons that compose the beam initiate
its journey in the Proton Source at Fermilab, where a pre-accelerator process
accelerates the protons from 0 to 750KeV through a RFQ cavity. After this, the
ions are injected into the LINAC.
4.3.4 The Booster and the formation of Buckets and Batchs
The linear accelerator (LINAC) accelerate H − -ions from 750KeV to 400 MeV
across a set of cavities that operate at a resonant frequency of 804.96 MHz. By
using magnets, the LINAC is able to provide the 400MeV protons to the Booster.
There are two RF frequency used in the LINAC: 201.24 MHz and 804.96 MHz.
The later value is the RF frequency that goes into the Booster during the injec-
tion of protons that has to last the exact value of the revolution period of Booster
(2.2µs).
After injection is complete, the Booster’s RF accelerates the beam from 400MeV
to 8GeV, but with different RF frequency comparing to the LINAC. The process of
re-bunching the beam according to the new RF frequency is called paraphasing.
The Booster is a synchrotron made of 19 RF stations, that change the frequency
from 37.8 MHz to 52.8 MHz as the beam revolution as the beam period is reduced
from 2.2µs (injection) to 1.6µs (extraction)6 . So the buckets, will have a length in
time of 1/52.8M Hz = 18.9ns. The harmonic number for the Booster is 84, the
value of buckets that circulate along the Booster. A booster batch is composed
of this 84 buckets that goes into th extraction process.
The extraction is done by injecting the protons into the MI-8 beamline that
delivers the proton beam to the Main Injector, the BNB beamline, the Recycler or
the Booster dump.
6
High energy particle in synchrotrons will have higher values of revolution period, and low
energy particles; will have lower values of revolution periods
Figure 4.6: How is the booster bacth formed and the value of it.
4.3.5 Main Injector: formation of MI Cycle, Resonance Extrac-

tion and Spill frequency
After the Booster, the beam proton is injected to the Main Injector where it ramps
up the kinetic energy from 8GeV to 120 GeV. Made of 20 RF cavities, during
the acceleration the frequency sweep from 52.8 MHz to 53.1 MHz, allowing the
transfer from bucket-to-bucket from the Booster. While the protons are circulating
the time beam structure remains the same, and the buckets slightly shrink in time
to 1/53.1M Hz = 18.8ns.
As the same with the Booster, due to the length of the circumference of Main
Injector (3319.4 m), 7 booster batches are injected during a time of 11.2µs. This
is the value of the MI Cycle, the time that takes a particle to go around all the
machine.
After reaching the 120GeV energy, the extraction process is done in order
to deliver beam to the different beamline that provide beam to experiments like
NuMI, or which is the case for this thesis, to the Test Beam (MTest) Facility where
the data has been taken.
The extraction of the beam that is use in the MTest is made using the process
call resonance extraction or spill duration that allow us to have a long low-
intensity pulse, in other words the beam is shaved off on every rotation the beam
makes around the machine, which in our case is 375000 MI cycle giving the
length in time of 4.2s each 60s. Typically, only the first batch will have particles in
it and the process is repeated around each minute. This interval of time is called
Spill frequency
Figure 4.7: Diagram showing the MI’s cycle formation, spill duration and spill fre-
quency.
Chapter 5
Tools for Data Analysis
In the chapter (1) we describe the weak interactions interactions and how the
neutrino’s mass change the experimental neutrino physics research, shifting the
focus into neutrino mass oscillations and the interaction of neutrinos with matter
experiments (1.2.1, 2.2).
And since MINERvA (3) started in 2009, its construction was made in order
to have fully functional detector of neutrino interactions in the low and medium
energy (with publications like B.6). The success of this kind of analysis rely on
the ability to remove the systematic errors and be able to reduce them in order to
get the true constructed energy of the incoming neutrino by improving the models.
That is why exists the MINERvA Test Beam experiment (3.2.1) which is cur-
rently studying the calorimetric detector response in the medium energy. In this
experiment is important to know the energy and type of the incoming particles by
knowing as best as we can the characteristics of the beam. Chapter 4 explain
why the structure of the beam has three scales (bucket, batch and spill) and how
the length in time of those structure are set up. The appendix C backs up this
analysis with a more theoretical description of the Radio Frequency cavities, the
key elements in the acceleration of the beam.
In this chapter, we outline the results of the analysis of the time structure of the
Test Beam program for medium energies. In the first section we describe some
general concepts about Statistical Analysis 5.1 which is mostly descriptive statis-
tics and the use of the statistical package 5.1.1 ROOT. In the sec 5.2 we describe
71
CHAPTER 5. TOOLS FOR DATA ANALYSIS 72
the data taken in the experiment, in 5.3 the developed tool (named TbTaTool) and
in sec. 5.3.3 the implementation of these tools into the data.
Finally the results are shown in 6.1 and 6.2, while the conclusions in 7 and
further work are describe in ??.
5.1 General concepts of Statistical Analysis
Statistics is the science of drawing conclusion from data, data are measurements
of some quality or quantity of the world, a value that can be obtain when we ap-
ply a tool to an observable variable, for example: a ruler marked off in inches is a
measurement tool for measurement the variable length of an object. The object
can be related with this measurement, and formally it can be defined as logic set
of rules and steps that tell us how to apply the measurement scale to an observ-
able variable.
The data is composed of only approximate measurement from the true mea-
surement. If we increment the sensitivity in our instruments, we can improve on
many decimal places but there is a limit to even the most powerful instruments.
A systematic errors of measurements, result from weakness in the mea-
surement procedure that produce distortion making the measurement always too
large or too small, while the random errors are produced by random stochastic
variations in the measurement. Two familiar concepts, accuracy and precision,
are deeply related with the type of errors.
Accuracy of a measurement refers to how close the measurement made is to
the true measurement, depending of the sensitivity of the measuring instrument
and the presence of errors in the process of measurement, particularly system-
atic errors. Precision of a measurement deals with how identical measurements
are able to reproduce similar values.
The tools of statistics are divided by two big divisions: descriptive statistics
and inferential statistics. While descriptive statistics ordered the data from the
measurements in order to gain insights about it, inferential statistics, infer prop-
erties of a population from a random sample. Histograms are great tools for data
visualization, compose by bars with their areas’ being the fraction of data in each
class interval. Finally apart from the usual measurements of central tendency, it is
important to remember the definition of the standardd deviation, which measure
the average distance from the data to their mean (or the RMS of the deviations
of the data from their mean).
Figure 5.1: A standardd histogram with box plots generate in ROOT.
A way to resume all the statistical information from a set of measurements is

throught the five-number summary or box plot, which is floating-rectangle graph
constructed horizontaly with the measurement scale along the X axis, or respec-
tively in the Y axis. The graph displayed x̄±s and the range. The rectangle’s ends
display the first (25%) and thrid quartile (75%) along the measurement scale. The
horizontal line through the box is the mean.
In the fig. 5.4, we show a histogram formed by different measurements and
box plots (blue boxes) for each of them.
5.1.1 ROOT: Data Analysis Framework
”ROOT is a cross -plataform C++ framework for pentabyte data storage for statis-
tical analysis and visualization”[4], developed in order to manage large amounts
of data in a efficient way. For example, the expected data amount produce by
LHC is (1PB = 10E6 Gb) per year1 . ROOT is cross-plataform since implemen-
tation of ROOT in C++ or python allow the user to take advantages of the inner
characteristics of these language programs.
The hierchacy of the programs follows the model of trees and branches. A
ROOT file contains all the data saved in different chains. A chain is a col-
lection of Trees, a Tree can have many branches and each branch has many
leafs (this is call a n-tuple). Each leafs are variables that we read through each
event. Since these events produced by the experiment are statistical indepen-
dent, they all have the same data structure. That is why ROOT has a hierarchi-
cal object-oriented database with plentyful of package for statistical analysis and
high-performance data processing.
(a) A tree named tree1, and all the (b) Data recorded from the data readout aux-
variables that where recorded iliary system, a n-tuple.
(c) How the data is recorded for each event. The rows correspond to the values and the columns to
the variables. ROOT reads the information acording to branches, not as events. A Ntuple is storage
in terms of events (horizontal).
Figure 5.2: Structure of the data in a ROOT file.
Having the same data structure with many events, the analysis are done run-
1
In case of MINERvA, the NuMI low-energy data size is around ∼ 3 GeV with about half of
those data are beam spill data and half are calibration data [29].
ning macros (a collection of well defined steps) which automates the analysis
performed for all the events. Actions like fitting a histogram, plot histograms and
made correlations graphics, include bar errors and present the data in more com-
prehensive visual format is done with few lines of code in ROOT. However, the
curve of learning is higher since the C++ and specific sintax requirements are
used by ROOT.
In the next section we will explain some concepts of statistical analysis while,
showing some commands in ROOT that allows us to work with them.
Fitting a histogram
Fitting a histogram allow us to get an approximation of the parameters that char-

acterize the curve like the mean or the standard deviation. This can be achieve
in ROOT’s graphic mode or through algorithms that includes them.
Figure 5.3: Fitting a histogram made with three lines of code.
[ root ] TF1 func ( " mydoublegauss " ," gaus (0)+ gaus (3) " )
[ root ] func . SetParameters (5. ,5. ,1. ,1. ,10. ,1.)
[ root ] hist1 . Fit ( " gaus " )
[ root ] hist2 . Fit ( mydoublegaus )
Running a Macro
A macro is a set of well defined steps that are going to be apply to each event of
the variables that we are using to analyse.
void AnalyzeSpill :: Loop ()

{
if ( fChain == 0) return ;
Long64_t nentries = fChain - > GetEntries ();
Long64_t nbytes = 0 , nb = 0;
" LINE1 " // == Begin of the Loop ==
for ( Long64_t jentry =0; jentry < nentries ; jentry ++) {
Long64_t ientry = LoadTree ( jentry );
if ( ientry < 0) break ;
nb = fChain - > GetEntry ( jentry ); nbytes += nb ;
" LINE2 " // == The Loop over all the events goes here ==
// == The cuts are apply here ==
if ( In_spill > 0.5 ){
Time_spill = ( Double_t ) Time -1429838450;
}
else if ( Spill_number == 1 ){
" LINE3 "
// == Assign values to the variables
Time_spill_1 = ( Double_t ) Time -1429838450;
tree_spill - > Fill ();
}
}
f_spill . Write ();
In LINE 1 we are initiation the ”loop”, the set of steps that will be repetead for
all the events that pass the cuts or condtions. From the LINE 2 the coditions are
placed using the if conditional if( condition ) steps . Finally in LINE 3 we calculate
the variables with the filtered sample. In our case we were subtraction a fixed
value to the timestamp and then filling the result into the root file.
Simple Analysis
Lets assume that we have a set of data with three variables: energy of the beam
(ebeam), final momenta in the x,y and z direction (px, py and pz). In this example,
a particle is traveling in a positive direction before it is deflected by a material at
a distance zv.
We can create a histogram of one of the variables (a), or plot two variables at
once (b), create a scatterplot observing the correlations between two variables or
apply conditions (c -cuts) to the variables in order to get insights (d) (fig. 5.4)2 .
(a) tree1→Draw(”ebeam”) (b) tree1→Draw(”ebeam:px”)
(c) tree1→Draw(”px:py”) (d) tree1→Draw(”zv”,”zv<20”);
Figure 5.4: Structure of the data in a ROOT file. The comments below the plots
are the ROOT commands.
2
Due to the process of learning ROOT during the development of these tools, a full manual
has been written and can be accessed here https://www.dropbox.com/s/k64vh4k52smlqmm/
manual_root.pdf?dl=0.
5.2 Description of the Spill scale in the beam at

MTest
In the former section, I have described how to fit, running a Macro and make a
simple Analysis with just one line. However in order to gain insights about the
time structure of the beam thought the TB data, the code is more complicated.
Since we are measuring a physical variable with an instrument, it is important
to know how the process of measurement is done and what are we actually mea-
suring. This will help us to know not only the sources of systematic errors but
also to consider the necesarry filters or cuts in order to eliminate the background
or data that it is not important to the analysis. In this thesis it has been important
to know how the detector and auxiliary systems have been designed, the read-
out process of taking data, the resolutions timing and the characteristics from the
input. I will explain more on these points.
The auxiliary system of the Test Beam (sec. 3.2.3) has been desing in order
to maxime the number of particles of know momentum and type that will be de-
tected. This impose some specific constrains to the readout system and DAQ,
with the most important one that the system must discard events that have not
pass the ”trigger criteria”. As mentioned in the sec. 3.2.4, the triggers conditions
avoid of taking data for other events than the particles that we requested to the
AD. So we only see events during the spill, anything else is background and
can not be analyze.
As mentioned in the sec. 3.2.2 and 4.3, there is a specific structure in time
which depend of the RF frequency and the harmony number of the Main Injec-
tor synchrotron. Particles come in ”packages” (buckets, batchs or spills) and
outside these we do not have any particle in theory. The trigger criteria does not
allow us to see out side the spill, all our results can not prove that the hole struc-
ture advertised from the AD is correct, it only can say that the structure during
the delivery of particles can be or not correct.
As it has been stated in the previous sections sec. 3.2.1, it is of real impor-
tance the characterization of the secondary beam that we recieve as an input for
the Test Beam MINERνA’s detector. One of this variables is the timestamp in
which we receive the particles (π ± , e± or µ± ) without knowing their identity. This
thesis has accomplished the study for one scale: the spill scale. In other words,
we will describe the time structure of the beam for the first scale. As we recall,
the spill is the length in time that the Main Injector (4.1.1) deliver particles into
the MTest at fermilab, the oficial time: 4.2s is the results of the slow extraction
process from the main ring sec. 4.3.5.
5.2.1 Description of the data
During the data takeout, the TB detector has been modified in two mode: ECAL/H-
CAL and Tracker/SuperHCAL and data has been taken during three different pe-
riods of time during the year of 2015.
Each mode represent a different configuration of numbers of planes for the
ECAL and HCAL regions mentioned before in the sec. 3.2.2. The followig table
resume them:
Table 5.1: Configurations of the Test Beam detector and the type of particles that
contain them.
Run Energy Type Configuration Date
Run 1 1.55,1.77,2,3,4,6,8 GeV π ± ,e± ECAL/HCAL 6-21 April
Run 2 4,6,8,9,10,16 GeV π± Tracker/SuperHCAL 23-30 April
Run 3 2,3,4,5,8 GeV e± Tracker/SuperHCAL during June
However, the ”Runs” category has more impact in the PID analysis and not
in the timing data which do not differentiate between different Runs or identity of
particles. In this thesis we will present the results vs energy and polarities.
It can lead to some confusion but, Run3 or Run2 are refering to the time and
configuration of the detector in which the data was taken. However, the data
process has been ordered using some similar words: run, subrun and gate. In
the fig. 5.5 it can be seen a diagram of how the data have been taken. A gate is
the internal variable of the CAMAC TDC 3377 which roughly speaking correspond
to one event that has pass the trigger criteria. A subrun is the collection of gates
(d), usually fixed in 1000. And a run (a,b and c) is a collection 74 subruns. Each
run contain data from an specific value of energy and polarity of an expected
type of particle and it is customary that one subrun will have 6 spills in the case
of electrons which corresponds to the variable Spill number.
Figure 5.5: Diagram of how the data is ordered in the ROOT file. The variable
shown is Time of readout. (a) and (c) are the energies that Run 1 contains for
electrons and pions in both polarities (all the energies are presented in tab. 5.1).
(b) Run 2 and Run 3 contain onlu electrons and pions. (d) Each Run has a
number of subruns. (e) Each subrun constains the the data for 1000 gates or
events. (f) One of the variables is the T ime variable as shown in this figure.
As can be seen, the Run 2 and Run 3 corresponds to only pions or electrons.
If those runs have only electrons or pions, a mix of them or if there muons, is a
task for the PID analysis made by A. Zegarra. We only use the Run2 and Run3
category as the data that contains only pions and electrons, respectively.
5.3 Development and Features of the Time Analy-

sis Tool (TbTaTool)
The analysis has been made with the data from Run2 (only pions) and Run3 (only
electrons) for the following time variable: T ime, the timestamp3 in which an event
3
The timestamp is the time in which and event was recorded. The CAMAC TDC 3377 uses
the unixtime which is defined as the numbers of seconds that has passed since 1/1/1970.
comes into the detector readout by the CAMAC 3377, regarding their type.
The variable that we used in order to eliminate the background were: In spill >
0.5 which assure us that we where looking particles inside the spill. This signal
is provided by the Acceleration Division. With this variable and restiction, we are
able to study:
1. The time profile of a variable and its reliance with a category-variable.
2. The study of any time structural difference in the variable and its depen-
dency with a category-variable by constructing ”time slices”.
3. Values of any function of the time and its reliance with a category-variable.
Some words about the general terms that I am using here. By category-
variable I am refering to any variable that in the process of measurement help us
to characterize the data that we are taking. For example: energy, spill number,
polarity, machine that took the data, etc. The use of this variables allow us to
look deep into how different parts of the machine are reading out the data, and in
terms of the analysis, in the case of category-variable = energy or polarity it allow
us to see any difference regarding those variables in the time structure.
And with ”any function of time” we are talking about any variable that is con-
structed with some calculation regarding time. For example, in our case we calcu-
late the Spill duration (the time that takes the resonance extraction) and the Spill
frequency (4.3.5). In the fig. 5.12 it is shown how is calculated the Spill duration
and the Spill f requency.
However, the TbTaTool need two parameters: the numbers of spills and a zero
point in time. The numbers of spills are important since the gates are ordered into
spills. For pions and electrons the number was not fixed, by choosing the value
of this parameter we are rejecting a percentage of particles to be analyse.
In the case of the reference point, it is needed in order to study the time
structural difference, this means the point in which we set as our zero point for
each subrun. The following section describe how the values were set up.
5.3.1 Election of a reference point in time
We define the time profile as the ”slices of time” of the beam that corresponds
to different Spill number, calculated considering the time of the first event of this
slice. But the calculation is made using a point of reference during each subrun.
Since the data only have an internal category-variable that order them, but not
a point in time where we can start making slices of time profile, it is needed an
external reference point.
If we use an internal point of reference, the time profile will be bias for a couple
of seconds (2-3) as can be seen in the following figure 5.6.
(a) Diagram of the three points of reference that can be used in the calcu-
lation of the time profile.
(b) Time profile for one subrun considering the beginning of the ROOT file as
the the reference point (2 in the above figure).
(c) Same as (b) but now considering (3) as the point of reference.
Figure 5.6: Bias in the time profile by using an internal reference point measure in
seconds. As can be seen, the (b) elecction allow us to reduce the lost of events
between the zero point and the first event.
On the other hand we have the $39 signal, which is the timestamp that is
associate with the time after the start of the super cycle, which is start-of-spill.
The code E.4 match each subrun with the closest start-of-spill signal and in order
to use it.
(a) Interval between the $39 and the kick-off the (b) Interval between the $39 signal and the be-
subrun ginning of the root file for Pions.
(c) Same as (a) but for data sets containing (d) Same as (b) but for same electron data sets.
electrons during June.
Figure 5.7: Distribution of the values of the interval between the first event
recorded (kickoff od the spill) and two reference points: the $39 signal and the
beginning of the subrun.
The interval between the first event and the beginning of the subrun is
around 33.63 seconds, half of the MI spill duration (fig. 5.7 (a)). In the case
of the time interval between the beginning of the spill - $39 shows that the
mean difference is around 0.16s (fig. 5.7 (b)). Less than a second and 210 times
smaller. That is why the tool use $39 as the reference point in the calculations of
the time profile. In the fig. 5.6 (b) and (c), it can be seen than by changing the
reference point we reduce the gap from 2-3 sto 0.5s.
The meaning of 0.5s will be explain in the Conclusions, since it is deeply
related with the MI’s Spill Frequency. In sec. E.3.1 and D.1.8 can be found part
of code and the documentation of the subroutine that make the match between
the first event of a subrun and the closest $39 signal from AD.
All the values of time were given in unixtime. The conversion into human-
readable time and vice versa, was made through a small python code (E.3.2). In
the fig. 5.8 it is shown part of the excel file with the timestamps in central US time
of the signal $39.
Figure 5.8: File that contains dates from the signal $39 as a reference point.
5.3.2 Number of spills
I have mentioned that category-variables allow us to classify the data while we

are taking them or to know how data is been recording respect the different part-
s/systems of the machine. One of this variables in our data is the Spill number.
The numbers of spills for each subrun is not a fixed number, instead it depends
if the number of events (1000) were achieved. The importance of knowing the
number of spills for both runs rely on the need to know as an input in know how
many spills will TbTaTool cut the spill in order to construct the time profile.
For Run 2, the mean value of spills is 6.287, while Run 3 has 9, 86 as the
number of spills during a subrun. The numbers of spills for Run 2 and Run 3 are
shown in fig. 5.9. In (a) the distribution of spills for Pions and in (b) for Electrons.
The plots are classify in positive and negative polarities, and the stability of spill
number during the data taking.
There are cases in which we have more than 15 spills per subrun. The Tb-
TaTool has fixed numbers for the number of spills. For Run 2, 6, while for Run 3:
10. This election do not affect greatly the statistics with which we calculate the
other features. In Table 5.2 can be observed the total numbers of events for both
situations.
If we consider all the spills, or if we choose 6 (as in the case of Run2) or 10
(Run 3). The last line is the percentage of electrons over pions.
(a) Run 2 only π ± . (1) and (2) are the distribution of the numbers of spill versus
the polarity. They have the same mean value. (3) Shows the stability of the
numbers of spills during all the data taking.
(b) Run 3 only e± . For (4), (5) and (6) same meaning as in (a). The mean
number of spills in this case is 10.
Figure 5.9: Number of spills for Run 2 and Run 3.
Table 5.2: Number of events for Run2 and Run 3 considering all the spills, equal
to 10, 6 or 2 spills per subrun.
Run All spills ≤ 10 spills ≤ 6 spills ≤ 2 spills
Run 2 122285 122285 116782 39111

π± 100% 100% 95.50% 33.49%
Run 3 206432 195383 119648 40702
e± 100% 94.65% 57.96% 19.72%
% 59.24% 62.59% 97.60% 96.06%
In the case of Run 2 by using only 6 spills per subrun we have 95.50% of all
the events. For electrons, Run 3, 6 is not a good election. And by using 10 spills
we are analysing the 94% of all the events.
5.3.3 Implementation of the tool for the Test Beam data
The fig. 5.12 is a diagram of how the TbTaTool developed calculate the time
profile and any function in time. This part of the tool pile up or stack the different
spills for a subrun, and loop over all the subruns according with the condition
imposed (for different energies, polarities or type of particle).
In each iteration within a subrun, the value of one Spill Frequency (or a mul-
tiple of it) is removed to the timestamp of the event, in order to set all the first
events around the time zero. The value of the Spill Frequency was provided by
the Main Injector according to the fig. 5.10. Before 15/5/5, 60.5246 ± 0.0968 and
after; 60.2415 ± 0.1929.
Figure 5.10: MI’s Spill frequency during May 2015. Infor provided by A.D.
Figure 5.11: How the TbTaTool produced the time profile. In (a) we show a diagram
of how the spills come in time and are arranged by the variable Spilln umber. As it
is shown, the point (2) and (3) were used as a reference point. The TAT ”cut” the
spills (b) independently from one another and then, stack them in one plot (d).
The plot can be produced for different energies, polarities or type of particles (c).
Spill Duration is calculated within one spill, the code take the first and last
event for that spill, subtract them and recorded in a root file. The same steps
are done for all the subruns and energies. Spill Frequency is calculated, by
taking the first event from two adjacents spills, and calculating the interval of time
between those two points. A simple version can be found in E.2 and in the github
repository 4 .
4
https://github.com/gsalazarq/TbTaTool/tree/master/TbTaTool
Figure 5.12: From (a) to (e) apply to one subrun. (b) the different points of refer-
ence in order to get the time profile. (c) and (d) explain how the spill duration and
the spill frequency are calculated. When the calculation for one subrun is done,
the next is calculated until the last sunrun. (f) and (g) describe how all of them
are stacked and plot (i) and (j) according to different variables (h).
Chapter 6
Results
6.1 Time profile of the MTest Beam
We present the time profile (slices of time) of the beam for the data sets containg
pions (Run2) and electrons (Run3)1 .
Time profile of the MTest Beam for Run2
The following plot 6.1 and 6.2 shows the time profile of the Pions in the interval
of a Spill for all the energies and polarities. In the bottom part of the plot the
length of the spill duration is showed. As we can see, the spills correctly are in
the window of time of 4.2s except for the events that are before the zero point.
1
In this chapter we use ± for different energies since we have data for electrons and positrons,
as well with positive pions and negative pions. The signs describe the charge of the beam deliv-
ered.
89
CHAPTER 6. RESULTS 90
Figure 6.1: Time profile for Run2 ±4, ±6 and ±8 GeV (pions).
Figure 6.2: Time profile for Run2 ±9, ±10 and 16GeV (pions).
Time profile of the MTest Beam for Run3
Figure 6.3: Time profile for Run3 ±2, ±4, ±8, 3 and 5GeV (electrons)
6.2 Main Injector’s Spill Frequency for Run2 and

Run3
The resonance extraction is made each minute according with AD, in order to get
the exact value the Spill Frequency has been calculated for Run2 and Run3 6.4.
(a) Resonance extraction for data set (b) Same as (a) but only for positive and
containing only pions for all the energies negative energies.
(c) Resonance extraction for data set (d) Same as (c) but only for positive and
containing only electrons for all the en- negative energies.
ergies
Figure 6.4: Spill frequency for Pions (a,b) and Electrons (c,d). Pions have energies
of 4, 6, 8, -8, 9, -9, 10 and 16 GeV, and electrons 2, -2, 4, -4, 8, -8, 3 and 5 GeV
For Pions, the rate of spills arraving to the detector is correctly calculate in
60.53 while the value provided correspond to before May 5. But there are clearly
two bumps on both sides of the principal spike in 6.4 (a b). This situation is not
present in the Run 3 electrons, where there are not bumps and the average value
is 60.21.
6.3 Main Injector’s Spill duration at MTest for Pions

and Electrons
We can see here the distribution of the spill duration fig. (A.13) for both Run2 and
Run3 according to the different energies (left) and polarity (right). In the first case
(a) and (b) the average value is 3.85s (all energies) and 3.81s (both polarities).
However, in the case of Run3, both histograms show the same value 3.63s.
(a) Spill duration for data set containing (b) Same as (a) but considering only
only pions for all energies. their polarity.
(c) Spill duration for data set containing (d) Same as (c) but considering only their
only pions for all energies. polarity.
Figure 6.5: Spill duration for all Runs (a,b) and Electrons (c,d). Pions have ener-
gies of 4, 6, 8, -8, 9, -9, 10 and 16 GeV, and electrons 2, -2, 4, -4, 8, -8, 3 and 5
GeV
In the tables of the fig. 6.6, we show all the values for the different variables
calculated:
• Delta time between the kickoff of the spill and the e39 signal.
• Delta time between the beginning of the subrun of the spill and the e39
signal.
• the MI’s Spill duration

• the MI’s Spill frequency
(a) Results for Run2
(b) Results for Run3
Figure 6.6: Errors against the oficial values of the MI’s frequency spill and MI’s
spill duration.
Chapter 7
Conclusions
There were three question that sowed the foundational idea of this thesis and
the development of a MINERvA Test Beam Time Analysis Tool (TbTaTool). As
mentioned in 3.2.1, a more comprehensive knowledge of the incoming particles
into the Test Beam detector is a requirement for improving the models detection
of final particle states of neutrino interactions.
• When a specific type of particle is coming in the beam timing?
• How frequently do we have (in the Test Beam detector) a π, e in one buck-
et/batch/spill?
• How can we use TB2 data to say something about the beam? Do we have
any restriction?
Saying that, our approach has been based in the development of a tool in
ROOT, python and C++ that it can help us to monitor the process of taking
data considering the constrains of how the detection systems was set up and
of course, the physics of the measurement itself.
Some words need to be say in order to contextualize and give a clear view
on these questions. This thesis is about study the timing of events in the detec-
tor, assuming that or with a further checking process, the identity of the income
particles. In order to successfully answer this, the TbTaTool allowed us to
study any time-variable defined by the user independently of the code. This
96
CHAPTER 7. CONCLUSIONS 97
means that it can be defined a priori in the taking data design process or a pos-
teriori inside the code as a function of time.
For example, since the time structure of the beam (chap. 4) has been defined
by others, we do not control it and the system of detection has to be defined
matching our physics needs and external constrains. That is why we are present-
ing results that were inside the spill (results secs. 6.1, 6.2 and 6.3) because
we want events with particles incoming the detector. Grab timing information by
studying the background of particles ”when we are not receiving particles” is not
possible in the current set up of the experiment. Knowing this, one success of this
work is that the TbTaTool can be applied to any case inside the spill or outside the
spill, any special difference is solved in how the input ROOT files are produced.
Needless to say, the TbTaTool does allows us to answer when a particle is com-
ing in the beam timing. Specially for the spill scale, successfully fulfilling the first
of three scales that define the time structure of the beam.
Furthermore, since the readout time is concentrated in the CAMAC TDC 377
(sec. 3.2.3), the conversion from analog signal to a digital one, resolution in tim-
ing of the different instruments (ToF, Veto, Cerenkov and Wire Chamber) impose
some limits in the analysis if we need to go further from nanoseconds, more work
need to be done in this area, but we are confident to be able to go into the batch
scale.
The third question express our curiosity of knowing if the TB2 data itself was
enough to study the time profile of the beam. We have a mixed answer. We do
study the spill and batch scale, but still we need a reference point in time. That is
why in the sec. 5.3.1 we needed to choose the closests e39 point to the kick-off
of th subrun in order to set our zero poin of reference.
The histograms showed in the 5.7 show that the election of the kickoff of the
spill above the beginning of the run is a correct election, since the gap between
boths is reduce to 0.16s, and not the 33.63 if we used the former. In fig. 5.6 we
pointed that there is a 0.5s of gap in the profile histograms. A related event is that
in fig. 6.1, 6.2 and 6.3 there are events before the zero point.
The reason is that, in order to stack the slices of time, we subtract a fixed value
of MI’s frequency spill according to the average value provided by Accelerator
Division as can be seen in the fig. 5.10 and as can be seen in the distribution of
values (fig. 6.4), the spill numbers it is not fixed, that is why when the calculation
is done an excess of deficit is always present. Further work needs to be done,
including a subroutine which match the value of the MI’s Spill frequency according
to the date in which the data was taken.
Also, as a footnote, we chose to fix the spill number (6 for Run2 and 11 for
Run3) without excluding events that can affect the statistics, showing in tab. 5.2
that with this election we have calculated up to 95% of the events (95.50% for
Run2 and 97.60% for Run3).
The implementation of the tool for the Time Profile (6.1), the MI’s Spill Du-
ration (6.3) and the MI’s Spill Frequency (6.2)demostrate that the tool is working
and giving us the distribution of value for this variables. In the tables 6.6 we show
those values. For the MI’s Frequency spill Run 2, we have a difference of −0.01%
with the nominal value by AD, while for Run3, is 0.06%. In the case of the MI’s
Spill duration, for Run2 we have a difference of 9.34% and Run 3 of 13.57%.
Finally we state that the development of the TbTaTool is important because
it has allow us to use the TbTaTool in other context and experiments, not
necessarily in the fields of particle physics. I will go deeper in this conclusion.
Science needs experiments which use instruments to make the measure-
ments and since we need to repeat a set of well defined steps in order to measure
a physical property from the object we are studying, computers are used: they
offer predictability in the measurements. This process is called Data acquisition,
a process that measure real world physical conditions and converting them into
digital signals that can be manipulated by a computer. Data acquisition usually
are abbreviated by the acronyms DAS or DAQ1 .
LabVIEW is one of the most well know system-design platform and develop-
ment environment for DAQ, that interface the sensors with the computer. In the
field of particle physics, DAQs use ROOT in the real-time visualization of the mea-
surement process. TbTaTool can be extended to this goal and be implemenented
1
In the chap. B I have described all the task that been a Detector Expert for the MINERvA
experiment involves. This training is basically a training deeply related with the DAQ systems of
the experiment.
in other experiments since the code is independent of the production of the data.
We are confident to build a real-time visualization tool experiment-independent.
Appendix A
Auxiliary plots
A.1 Spill frequency for only Pions (Run 2)
(a) -4GeV (b) 4GeV
Figure A.1: Spill frequency for 4GeV Pions
(a) -6GeV (b) 6GeV
100
APPENDIX A. AUXILIARY PLOTS 101
(a) -8GeV (b) 8GeV
(a) -9GeV (b) 9GeV
(a) -10GeV (b) 10GeV

(a) 16GeV (b) All energies
Figure A.6: Spill frequency for 10GeV Pions and all the events
A.2 Spill frequency for Electrons (Run 3)
(a) 2GeV (b) -2GeV
Figure A.7: Spill frequency for 2GeV electrons
(a) 4GeV (b) -4GeV
Figure A.8: Spill frequency for 4GeV electrons
(a) 3GeV (b) 5GeV
Figure A.9: Spill frequency for 3GeV and 5GeV electrons

(a) 8GeV (b) -8GeV
(c) All energies
Figure A.10: Spill frequency for (±)8GeV and all energies electrons.
A.3 Spill duration for only Pions (Run 2)
Spill duration plots for data set containing electrons (Run 3). Some fits do not
work, showing that more statistics are needed for improving this part of the tool.
(a) 4GeV (b) -4GeV
(c) 6GeV (d) -6GeV
Figure A.11: Spill duration for data set containing only pions (Run2).
(a) 8GeV (b) -8GeV
(c) 9GeV (d) -9GeV
(a) 10GeV (b) -10GeV
(c) 16GeV
A.4 Spill duration for only Electrons (Run 3)
(a) 2GeV (b) -2GeV
(c) 3GeV (d) 5GeV
Figure A.14: Spill duration for electrons (Run3).
(a) 4GeV (b) -4GeV

(a) 8GeV (b) -8GeV
(c) All energies both polarities.

Appendix B
Participation in the MINERνA

experiment during 2015
The MINERνA experiment is collaboration-joint between many institutions and

persons. The participation in the experiment is done in many different ways: by
doing and analysis of neutrino cross-section usually as a part of a PhD program,
by been involved in different teams that the Main Detector needs in order to check
the quality of the data, calibration, reconstruction and/or DAQ; and by doing shifts,
which means take periods of time of 8 hours in which the person is in charge of
keeping the data readout system online and fix any problem that may happen.
In this section I will outline the activites and additional work that I performed
during my participation during in the MINERνA experiment.
B.1 Shifts
”Shifts” are the short name that describe a obligatory tasks that all the members
of the collaboration have to do in order maintain the Main Detector continuously
taking data. These shifts are 8 hours long at least three days on a week, accord-
ing with a timetable that include all the active members of the experiment.
But not only a shifter (the person how is in charge of the shift) has to maintain
the data-taking runs but also ”ensure that the DAQ is running whenever possi-
ble while the beam is on, monitor the electronics for failures and problems, and
109
APPENDIX B. PARTICIPATION IN THE MINERνA EXPERIMENT DURING 2015 110
stop runs and reset the electronics in the case of a failure, log any electronics
or software errors, perform basic data quality checking and calibration analysis,
develop tools to do this when necessary and understand how to handle common
errors in the data taking and monitoring procedures”. Internal documents of the
Collaboration.
In this section I will briefly resume elements of this process of taking shift.
which include log all the events in the MINERvA logbook as a complement of the
?? and 3.1.2. First, it is mandatory to have control of the MINERvA Run Control
fig. B.1 (a) and that all the voltage values in (b) for the Near MINOS DAQ are in
the accepted values.
Figure B.1: (a) Left. Run Control of the MINERvA Main Detector. (b) Right.
Near MINOS DAQ, not all information is useful for MINERvA, just the two bot-
tom columns.)
Then, it is important to check if the Main Detector is receiving beam from the
NuMI Beam, in fig. B.2 (a) and (c) can be seen plots that characterize the quality
of the beam, the interval between the last spill of particles (usually 1.3s) and the
amount of protons on target that the experiments are receiving.
From time to time it is needed to check the quality of the data that we are
taking. That is why the fig. B.2 (b) is an automatic generated checklist where the
shifter accept or deny that that particular run contain data of good quality.
There are other mandatory tasks, like: fill out the start shift form, phone MCR
that you are on shift, fill out the Rock Muon Check List, check the MINOS DCS
Figure B.2: (a) Up left. NuMI Beamline Status Display which shows the status of
the beam. (b) Up right. MINERvA Veto Wall control. (c) IF Beam Data Server A9
event monitoring. (d) Bottom left. MINERvA Quality Checklist. (e) Bottom right.
Live event display for neutrino interactions in the Main Detector.
Checklist and the MINOS OM Checklist Near at the middle of the shift and Sum-
mary Plots at the end of the shift. The shifts ends when the next shifter take
control of the Run Control.
B.2 Detector Expert Training
We have talked about shifts, as the process of monitoring the data readout in the
Main Detector. But as any machine, it will face problems during the data taking
that even a trained shifter will no be able to resolve. A Detector Expert is a person
who has more training and knowledge that will resolve these problems or at least
call the right person to do it. Time is critical, and neutrinos interactions passing
our detector without been detected means money spent without any result.
Figure B.3: Diagram of the MINERvA DAQ System. The Detector Expert training
include a deeper knowledge in these areas.
The Detector Expert Training included:
• Replacement of malfunction FEB reported by the shifter
• Turn on and off the Main Detector in order to remove some critical parts of
them
• DAQ Firmware update
• Use the RunControl and SlowControl in order to check the voltage or any
error. Usually a shifter is not allow to change parameters.
• Return to Service the DAQ after any random event
• Change, Replacement of FEB
B.3 PMT Cross Talk problem
One important element in the DAQ and Optical System (3.1.2) of the Main De-
tector is the PMT. The Hamamatsu Photonics PMT model number H8804MOD-2
was chosen in order to transform photons into photoelectrons using a low quan-
tum efficiency photosensor with a timing resolution of 5ns that avoid overlapping
events within a single spill of the NuMI beamline. The photosensor has a 8 × 8
array of pixels laid out on a 2cm2cm grid. The 64 pixels have a spectral response
between 300 - 650 nm and can work between -30 to 50 °C . B.4.
Each of the pixels correspond to one scintillator strip from the Main Detector,
having more than 320000 channels. All the channels have been carefully oriented
in one and only one pixel. The problem of ”Cross Talk” arise when for unknown
reasons the photons that reach one pixel also illuminate the one of the four other
pixels around it. The task in which I participated was to help measuring if two
pixels showed response when one of them was illuminated, i.e. measure the
cross talk. In (b) fig. B.4 it can be seen two peaks on different x positions, are a
hint that cross talk is happening there.
(a) Equipment use for checking the (b) Usual data that shows that there are a prob-
Cross Talk problem on PMTs. lem of Cross Talk.
(c) Fiber optic pattern weave within each PMT

Box. Image from [2].
Figure B.4: Cross Talk tasks

B.4 PMTs Testings
Also, I participate in the process of checking the quality of the PMTs that have
problems in the Main Detector. The fig. B.5 shows the equipment, which is a
replica of the DAQ system in the Main Detector. There are 8 CROCE (Chain
ReadOut Controller) connected to one custom VME module (Versa Module Eu-
ropa, a computer bus standard). Each CROCE, one in the image, can support at
least 10 FEB (Front End Board) which controls the PMT and digitalize the ana-
logic charge through the TripP Chips that is place on each FEB. Then the CROCE
is chained with the CRIM (the CROC Interface Module)which distributes timing to
up to four CROCEs (3.1.2).
The work did included the diagnostic of the bad PMTs, to removed from the
Main Detector and the confirmation if the problem was becaus bad quality HV
delivery, light leak or something else. The methodology included to use the Run
and Slow control, to main components of the DAQ MINERvA, change the high
voltage, frequency, looking for specific behaviour on a specific pixel (from the 64
that contains one PMT box).
Figure B.5: (a) Left. Some photos the Sillicon Detector Facility and the equipment
where the PMTs’ tests were done. (b) Right. Run and Slow Control and some
histograms which confirmed the presence of cross talk.
B.5 Data and MonteCarlo Rock Muon eye scanning
This task was designed in order to confirm if the conditions or cuts imposed on
the Main Detector’s algorithms actually removed the rock muons events from the
data samples, leaving the muons produced by neutrino interactions.
We were given two sets of data (around 500 events): real data and Monte
Carlo simulated data. According with the nature of the charged particle, the en-
ergy lost on a specific material is different for each one, as well as the visual trail
in the event viewer. For example fig. B.6 correspond to Monte Carlo simulated
data. In this event it can be seen one neutrino interaction, but a lot of activity in
the upstream part of the detector even thought eliminating crosstalk that usually
makes messy the vertex of interaction that includes other charged particles that
produces other vertex. By checking the trails, the time slices of interaction (two
events at 4041ns upstream and 4028ns) and the point of vertex we are able to
discard this events as a rock muon event, which in theory was one of them.
Figure B.6: Monte Carlo simulated Rock Muon interaction with the detector.
In fig. B.7, real data shows two events: one rock moun at 3800 ns comming
in front of the veto wall and a neutrino interaction at 3732 ns.
The confirmation that for the last event was a rock muon, as well as the for-
mer one was not, helped the collaboration to improve the cuts imposed in the
algorithms and clean the real data from rock muons.
(a) (b)
Figure B.7: Real interaction of Rock Muons with the detector.

B.6 Participation in published articles during 2015
(a) Physical Review Letters Vol 116 (2016) 081802.

Title: Measurement of Electron Neutrino Quasielastic and Quasielastic like
Scattering on Hydrocarbon at Eν = 3.6 GeV
Authors: G. Salazar, A. Zegarra, C. J. Solano Salinas MINERvA colaboration
In this paper, one of the most important analysis in the lower energy carried
out by MINERvA, the first direct measurement of electron-neutrino quasielastic
and quasielastic-like scattering on hydrocarbon in the few-GeV region of incident
neutrino energy.
The results presented (flux-integrated differential cross section in electron pro-
duction angle, electron energy and Q2 ) will help current and future oscillation ex-
periments improve the detection process of CP violation measurements in the
neutrino sector (which is done by precise measurements of νe(νe)
¯ appearance in
predominantly νµ(ν̄µ ) beams like NuMI at Fermilab).
To maximize the rate of neutrino interactions, these neutrino oscillation ex-
periments like NOvA, T2K and DUNE use detectors of heavy nuclei (e.g. carbon,
oxygen, argon). By comparing the observed energy spectrum distribution with the
predictions based on different oscillation hypotheses, confirm or deny oscillation
parameter values that constrains the mass and the amplitude of the oscillation.
This paper reports precise measurements of the charged-current quasielastic
(CCQE) interactions νe n → e− p and ν̄e p → e+ n on ¯ep → e +n) on nucleons in a
hydrocarbon target at an average e energy of 3.6 GeV, in order to enhance the
correct prediction of the observed energy spectrum for νe interactions.
(b) Physical Review Letters Vol 116 (2016) 081802
Title: Charged Pion Production in νµ Interactions on Hydrocarbon at Eν =
4.0 GeV
Authors: G. Salazar, A. Zegarra, C. J. Solano Salinas MINERvA colaboration
Current and future long baseline neutrino oscillation experiments rely in study
of charged pion production via charged-current ν interactions since the nuclear
medium (the detectors themself) impact in the production and propagation of

hadrons produced in neutrino-nucleus interactions. Cross sections distorsions
which are absent in scattering from free nucleons and affect event rates and final
state kinematics, are examples of the impact of nuclear targets.
T2K and MiniBooNE oscillation experiments rely on well understood the CCQE
reaction νe n → e− p and the reconstruction of the energy on those events, but
the presence of nuclear medium introduce distorsions on the reconstruction and
interpretation of the events, as mentioned before. For example if the charged-
current interaction produces a single π + , the if in the process νe N → e− pπ + , the
pion is absorbed by the target nucleus, the event will mimic the CCQE topol-
ogy. Reconstructed neutrino energy may be underestimated, leading a bias in
the measured oscillation parameters. It is expected that analysis in the medium
energy are been running, where the new improved models for selection of muon
and pion events will include the results of UNI master students.
Appendix C
Radiofrequency Cavities theory
C.1 Electromagnetic Waves
We begin with a review of electromagnetic waves in vacuum and matter propa-

gate throught RF cavities RF are made in order to store electromagnetic energy
and kick the particles, increase the kinetic energy conductors have the ability to
reflect all the waves wave guides allow the propagation of waves throught hollow
cylinders Resonant Cavities have ends, which reflect and create stading waves
the maximun efficiency is reach when Q is high and depende of the geometry
throught this section we will explain this concepts applied to RF cavities
C.1.1 Electromagnetic waves in Vacuum
A electromagnetic wave equation is a disturbance of the electric and magnetic

fields that propagates with a fixed shape and at a constant velocity. In a
region where there are no charges or currents, the Maxwell’s equations are:
∂B
(i)∇ · E = 0, (iii)∇ × E = −
∂t (C.1)
∂E
(ii)∇ · B = 0, (iv)∇ × B = µ0 0
∂t
as we know we can derive two 3D-wave equations for the electric and mag-
netic field, using eq. C.1 operator relations, after all the mathematics:
119
APPENDIX C. RADIOFREQUENCY CAVITIES THEORY 120
∂ 2E
∇2 E = µ0 0
∂ 2t (C.2)
∂ 2B
∇2 B = µ0 0 2
∂ t
Maxwell’s equations imply that empty space supports the propagation of elec-
p
tromagnetic waves at speed of 1/0 µ0 , since the contribution of Maxwell’s term
to Ampere’s law allow it.
Monochromatic Plane Waves
The most basic expresion for a wave is to be a monochromatic wave (where

we only include one frequency) and aslo a plane wave, where the direction of
the propagation will be perpendicular to the fields (transverse). Considering the
complex expression. These wave are transverse, from (i) in C.1:
¯ · Ē = 0 → ∇
∇ ¯ · (Eei(k·r−ωt) ) = 0
E(ik · r)ei(k·r−ωt) = 0 (C.3)
k · r = 0,
both fields E and B are perpendicular to each other. Also they are in phase:
k
B(r, t) = (k̂ × Ē), (C.4)
ω
and the monochromatic planes waves can be expressed as:
E(r, t) = E0 ei(k·r−ωt) n̂
1 1
B(r, t) = E0 ei(k·r−ωt) (k̂ × n) = k × E (C.5)
c c
where n̂ is the polarization of the electric field. We can have a number of

quantities as the energy density (energy per volume) stored in the EM wave, the
flux of energy (Poynting vector) and the density of momentum(℘):
1 1 0
u = (0 E02 + B02 ) →= E 2
2 µ0 2
S~ = 1 (Ē × B̄) →= cuk̂ (C.6)
µ0
The flux of energy over a time ∆t is th density of energy that passes throught
a volume in the time ∆t with velocity c. Particles are accelerated in RF systems
by the transversal TM or TE waves where energy is transfered to them.
C.1.2 Electromagnetic waves in Matter
RF cavities are not hollow, they use dielectrics for example in order to tune the
magnetic field inside them. That is why we describe quickly the EM waves in
matter. The EM formalism make simple the study this subject, by just replacing
0 , µ0 → , µ we get the former results.
∂B
(i)∇ · D = 0, (iii)∇ × E = −
∂t (C.7)
∂D
(ii)∇ · B = 0, (iv)∇ × H = µ0 0
∂t
with no charge and no currents, again we have the 3D equation but consider-
ing the linearity of the fields D̄ = Ē and µ1 B̄ = H̄, the explict expresion for the
electric and magnetic field are:
E(r, t) = E0 ei(k·r−ωt) n̂
1 1
B(r, t) = E0 ei(k·r−ωt) (k̂ × n) = k × E (C.8)
v v
The index of refraction is defined as:
r
µ c
n= = (C.9)
0 µ0 v
√
usually the µ0 ' µ so n ' r .
C.1.3 Electromagnetic Waves in Conductors
So far we know that EM in vacuum or matter accept monochromatic plane waves

that propagate along the space with a well define velocity. In both cases, when
there reach a surface, the EM wave will have a reflected, transmitted and in-
cident wave. EMW in conductors consider the existence of free charges and
currents, that we don’t control ρf and J = σE(which follows) the Ohm’s Law so
the Maxwell’s equation are:
1 ∂B
(i)∇ · E = ρf (iii)∇ × E = −
∂t (C.10)
∂E
(ii)∇ · B = 0, (iv)∇ × B = µ0 0 + µ0 Jf
∂t
working the math in the solution, we get two partial differential equation that
∂
resembles the 3D wave equation but with the add of the third term µ0 0 ∂t . Monochro-
matic waves are still solutions of theses equations but now k is complex. The
imaginary part represent the attenuation that the material exert over the wave.
Ē(r̄, t) = Ēe−κz ei(r̄·k̄−ωt) (C.11)
Reflection at a Conducting Surface
Consider that in the region of study, there are free charge ρf and free current
density Jf electric and magnetic fields in both boundaries (in the last section we
did not consider that). From the Mawxell’s equations, the boundaries conditions
are:
k k
Ē1 = Ē2 B̄1⊥ = B̄2⊥
1 k 1 k (C.12)
1 Ē1⊥ − 2 Ē2⊥ = σf B̄1 − B̄2 = K̄f × n̂
µ1 µ1
The surface will produce from an incident monochromatic, a reflected and
transmitted wave in defined ratios1 .
1
The ratios are:
1−β 2
EOR = ( )EOR , EOT = ( )EOR
1+β 1+β
µ1 v1
where β = µ2 v2 is a complex number (since the number propagation for a linear media is also
In the case of a perfect conductor β is infinite and the incident wave will be
totally reflected, with a −180 phase shift.
C.2 Guide Waves and Resonant Cavities
The generation and transmission of electromagnetic radiation involves metallic

structures with dimensions comparable to wavelengths (meters) that we are work-
ing with. Hollow metallic cylinders that produce the propagation or excitation of
electromagnetic waves are called wave guide. Maxwell’s equations for this case
are (we assume a perfect conductor):
∂B
(i)∇ · E = 0 (iii)∇ × E = −
∂t (C.13)
∂E
(ii)∇ · B = 0, (iv)∇ × B = µ
∂t
The cylinder is filled with a uniform nondissipative medium having diaelectric
constant and permeability µ, both E and B satisfy:
ω2 E

2
(∇ + µ 2 ) =0 (C.14)
c B
(a) A simple wave guide, with a specific transver- (b) Five-resonator triple-post waveguide band-
sal shape where the EM wave enters. pass filter made by Ferranti. Source Wikipedia
CC1.0.
Figure C.1: Examples of wave guides.
complex taking account the attenuation of the wave).

The general solution will include a the sinusoidal dependece e−iωt from the
monochromatic plane waves and a cylindrical symmetry:
E(x, y, z, t) = E(x, y)e±ikz−iωt

(C.15)
B(x, y, z, t) = B(x, y)e±ikz−iωt
± describe the appropiate combination for standing or travelling waves in the

z direction. The wave number k is unknow. Making explicit the z-dependence eq.
C.14 takes the form:
ω2

E
[∇2t 2
+ (µ 2 − k )] =0 (C.16)
c B
Manipulating the eq. C.13 to C.16, for the simple case of the x-component we
have:
∂B
∇×E=−
∂t
∂Ez ∂Ey ∂Ex ∂Ez ∂Ey ∂Ex
=( − )î + ( − )ĵ + ( − )k̂ (C.17)
∂y ∂z ∂z ∂x ∂x ∂y
= iωBx î + By ĵ + Bz k̂
with similar expression for Bx,y,z . Since ∂Ex /∂z = E0x ei(kz−ωt) = ikEx , after
doing the math we get for the particular case:
i ∂Ez ∂Bz
Ex = 2 2
(k +ω ) (C.18)
(ω/c) − k ∂x ∂y
and for the general case:
1 ∂Bz ω
Bt = ω2
[∇t ( ) + iµ k̂ × ∇t Ez ] (C.19)
(µ − k 2 )
c2
∂z c
1 ∂Ez ω
Et = 2 [∇t ( ) − iµ k̂ × ∇t Bz ] (C.20)
(µ ωc2 2
−k ) ∂z c
We assumed a perfect conductor in the surface of the wave guide so: E || = 0

and B ⊥ = 0 inside the material itself and the inner wall, or:
n · B = 0, n×E=0 (C.21)
the vanishing tangential electric field E at the surface means Ez |S = 0), while
∂Bz
the perpendicular component of B equal to zero implies: |
∂n S
= 0 where ∂/∂n
is the normal derivative at a point on the surface.
This two boundary conditions can not generally be satisfied simultaneously,
therefore we have two distinct categories of EMW that can exist in the wave guide:
a tranverse magnetic mode (TM) and a transverse electric mode (TE).
Transverse Magnetic (TM)
Bz = 0, everywhere. Boundary condition: Ez |S = 0
Transverse Electric (TE) (C.22)

∂Bz
Ez = 0, everywhere. Boundary condition: |S = 0
∂n
From eq. C.19 and C.20 we see that if we find the Ez or Bz we can deter-
mine the appropiate solutions for the two-dimesional transverse wave equation.
A special case when both conditions are met is call TEM mode, and in this case
√
the wave number k = µω/c
From eq. C.19 and C.20 and considering a dependece in z of eikz
TM waves:
ik µω
Et = ∇t Ez , Bt = k̂ × ∇t Ez
γ2 ck
TE waves:
ik ω
Et = ∇t Bz , Et = k̂ × ∇t Bz
γ2 ck
As mentioned before for TM mode it is enough to know Ez and for TE; Bz .

ψ = (Ez , Bz ) has to satisfies the two-dimesional wave equation C.14 for the z
component:
(∇2t + γ 2 )ψ = 0 (C.23)
with the boundary conditions C.22 correspondingly for TM(TE) waves.

∂ψ
ψ|S = 0 , or |S = 0 (C.24)
∂n
From this, it is clear that constant γ is not negative and will take a spectrum of
eigenvalues γλ2 , and the differents solutions for λ = 1, 2, 3, ... will be modes of the
guide. For a frequency ω the wave number is determined according to the values
that λ takes:
ω2
kλ2 = µ( − γλ2 ), (C.25)
c2
ωλ is defined as the cutoff frequency:
γλ
ωλ = c √ (C.26)
µ
and the wave number can be written as:
1√
q
kλ = µ ω 2 − ωλ2 (C.27)
c
it can be seen that, for ω > ωλ the wave number is real and the wave propa-
gates throught the wave guide. When it is not positive, kλ is imaginary, and it is
attenuated while it propagates.
C.2.1 Resonant Cavities
A special type of cylindrical wave guides with end faces are called cavities, as can
be seen in the fig. C.2. The cavity’s wall are taken to have infinite conductivity,
while the cavity is filled with a lossless diaelectric with constants µ, .
Now we have to take into account the full reflections on the ends of z as
a additional boundary conditions for eq. C.19 and C.19, which means that the
waves will be reflected, an appropiate choose for standing waves are:
ψ ∝ A sin kz + B cos kz (C.28)
where the wave number is k = p πd , and d the height of the cavity . For TM
fields the, Et = 0 for z = 0 and z = d means:
Figure C.2: A pillbox cavity. The lower mode frequency does not depend from the
height of the cavity.
pπz
Ez = ψ(x, y) cos( ) (C.29)
d
While for TW fields, Bz = 0 for z = 0 and z = d requires:
pπz
Bz = ψ(x, y) sin( ) (C.30)
d
The general solutions for Et and Bt , for the transverse fields are:
TM waves
pπ pπz
Et = − 2
sin( )∇t ψ
dγ d
iµ pπz
Bt = − 2 cos( )k̂ × ∇t ψ
cγ d
TM waves (C.31)
iω pπz
Et = − sin( )k̂ × ∇t ψ
cγ 2 d
iµ pπz
Bt = − 2 cos( )∇t ψ
cγ d
From C.25 the eigenvalue problem reads:
ω2 pπ 2
γ 2 = µ − ( ), (C.32)
c2 d
for each value of p the eigenvalue γ 2 λ determines an eigenfrequency of reso-
nance frequency ωλp and the resonance frequency ωλp is:
2 c2 2 pπ
ωλp = [λγ + ( )2 ] (C.33)
µ d
A practical resonant cavity is the right circular cylinder (pillbox). Choosing
cylinder coordinates in the TM mode the transverse equation has a solution that
use the Bessel functions and an angular dependence in the φ. For ψ = Ez and
with the boundary conditions Ez = 0 at ρ = R, the proposed solution:
ψ(ρ, φ) = Jm (γmn ρ)e±imφ (C.34)
where, the independent variable for the Bessel functions, has a factor that
depends proportionaly to the n-th root (Jm (x) = 0) and R, the inner radius of the
cylinder:
xmn
γmn = (C.35)
R
The resonance frequencies depend on three indexes one from the perodicity
of the cavity and two from Bessel’s solution It has the same structure.
r
c x2mn p2 π 2
ωmnp =√ + 2 (C.36)
µ R2 d
The explicit expresions for the lowest resonance frequency in TM mode (m=0,n=1,p=0)
are:
2.405
ω010 = √ c (C.37)
µ R
2.405ρ −iωt
Ez = E0 J0 ( )e
R (C.38)
√ 2.405 −iωt
Bφ = −i µE0 J1 ( )e
R
which are independent of d.
C.2.2 Power Losses in Cavity: Q of Cavity
Resonant cavities have definite field configuration for each resonance discrete
frequency of oscillation. Fields will not built up unless the exciting frequency
matches the resonance frequency in reality there is a narrow band of frequen-
cies around the eigenfrequencies where excitation occurs. a measure of the
sharpness of response of the cavity to external excitation is Q value of the cavity
defined as:
Stored energy
Q = ω0 (C.39)
Power loss
definition of Q the ratio of the stored energy and the power loss times per
cycle the frequency of excitation assuming ohmic losses, the behavior of the
stored energy is
dU ω0
=− U U (t) = U0 e−ω0 t/Q (C.40)
dt Q
for the stored energy the change of energy in time is proportional to the energy
stored at that moment. This time dependence implies that the electric field is also
damped
E(t) = E0 e−ω0 t/2Q e−iω0 t (C.41)
but accepting that there is no single frequency but a superposition of fre-

quency around ω = ω0 , the Fourier decomposition regarding the energy’s fre-
quency will be:
Z ∞
1
E(t) = √ E(ω)e−iωt dω
2π −∞
Z ∞
1 (C.42)
E(ω) = √ E0 e−ω0 t/2Q e−i(ω−ω0 )t dt
2π 0
And the frequency distribution for the energy in the cavity can be calculated
from
1
|E(ω)|2 ∝ (C.43)
(ω − ω0 )2 + (ω0 /2Q)2
Figure C.3: The resonance curve’s full width is equal to the central frequency ω0
dived by Q.
which has a Lorentz line shape shown in the fig. C.3 with a full width at half-
maximun equalto to ω0 /Q. The energy of oscillation in the cavity will follow the
resonant curve in the neighborhood of the particular resonant frequency. ∆ω,
the frequency separation between half-power points, so Q os the cavity can be
defined as:
ω0
Q= (C.44)
∆ω
having Q values of severald hundreds for microwave cavities.
C.3 ToF basic physics
The momentum can be determined if we know the time of flight and the path
lenght with the following equation,
1 L(m0 c2 )
p= √ (C.45)
c t2 c2 − L2
with an error propagation depending on time and lenght.
t2 c4 m2
σp2 = [L2 σt2 + t2 σL2 ] (C.46)
(t2 c2 − L2 )3
also it can be derived that, two particles with the same momentum but different
masses have different time of flight.
s s
L m21 m22
T1 − T2 = ( 1+ − 1+ ) (C.47)
c p2 p2
Appendix D
Documentation of Code
D.1 Subroutines
D.1.1 Numbers of events in one subrun
Command name: NumberEventsInSpill()

Description: This function calculate the numbers of events in one subrun or one
root file. With D.1.2 we can calculate the total number of events in a set of file,
including the numbers of events for all the data set that we are analyzing.
Table D.1: Calculation of the number of events in one subrun.
Input Description output Description
char* file_name path of the root file to be calculated values_NumberEventsInSpill[r] numbers of the r − th
spills in the input file
"file_name"
int value_NumberSpills fixed value of spill number in the root files.
for Run 2 (6)m for Run 3 (11)
int r number of subrun’s spill that will be calculated
D.1.2 Numbers of events for a Run
Command name: NumberEventsInSpillForRun()

Dependencies: D.1.1
Description: This function calculate all the events in a set of files, usually a ”run”.
A run is a collection of subruns (usually 74 of them). The total number of subruns
is defined by the tool input’s limits ( file_begin and file_end). Since there is
132
APPENDIX D. DOCUMENTATION OF CODE 133
no dependence in energy, this subroutine do not classify the results into energy
or polarity. Each subrun is an element of the array name_file[1000] now set to
be 1000 of file.
Table D.2: Calculation of number of events for a set of root files.

int file_begin first root file path to be calculated NumberEventsInSpillForRun.txt txt file with the number
of events for all the files
int file_end path of the last root file to be calculated
D.1.3 Extracting the timestamp of one subrun
Command name: BeginTimeOneSubRun()

Description: Calculate the timestamp of the subrun or root file, by getting the first
value that was recorded by the CAMAC TDC (that passes all the Veto conditions).
Table D.3: Getting the unix timestamp for one subrun.
Input Description Output: t_begin Description

char* file_name path of the root file to be use timestamp for the $file_name root file.
If it is not calculated the value is
set as t_begin = -404
int file_end path of the last root file to be calculated
D.1.4 Extracting the timestamp for the first event of all spills
inside one subrun
Command name: BeginTimeOfSpillsForOneSubRun()

Description: This subroutine calculate the timestamp for the first event of all spills
inside one subrun. The time is returned in Chicago local time.
Table D.4: Getting the unix timestamp for the beginning all the spills in one subrun.

char* file_name path of the root file to be use BeginTimeOfSpillsForOneSubRun[m] timestamp value of the beginning
int m Spill number to be used of the spill number m
Long64_t nentries numbers of entries in the root file
nentries=fChain->GetEntries()
D.1.5 Calculation of the Kick-Off of One Subrun
Command name: KickOffOneSubRun()

Description: This subroutine calculate timestamp of the first event inside the first
non-zero event spill within a subrun.
Table D.5: Return one timestamp for the real beginning of one subrun.

char* file_name path of the root file to be use value_KickOffOneSubRun Kick-off of the subrun
Long64_t nentries=0 numbers of entries in the root file
nentries=fChain->GetEntries()
D.1.6 Generation of txt file with timestamps of beginning of

the subrun
Command name: BeginTimeForRun()

Dependencies: D.1.5 and D.1.4
Description: Create a txt file with the timestamps for two cases. If the which_tbegin=0,
the function will record the timestamp of the beginning of the subrun using BeginTimeOneSubRun()
(D.1.4); if which_tbegin=0, the data saved will be the timestamp of the first event
in the subroot using KickOffOneSubRun() (D.1.5).
Table D.6: Function that create a txt file with beginning timestamps of root files.
int i_begin index of the first root file
int i_final index of the last root file
int which_tbegin= 0 if we use Run 2 (0) or Run 3 (1)
D.1.7 Matching the two closest timestamps
Command name: MatchTwoPoints()

Description:
From two set of points, this subroutine matches the closets ones. In this case,
e39time[k] is the set of points that we need to match with BeginTimeForRun()
time of the root file. Also, it gives the distances and both points in time.
Table D.7: Function that matches a set of two points.

int r index of the result value_MatchTwoPoints[0] e39time matched value
Double_t t_begin input time to be matched value_MatchTwoPoints[1] delta time for the matched values
value_MatchTwoPoints[2] tbegin as an input
D.1.8 Matching two points for a set of subruns during a data

run
Command name: MatchTwoPoints()

Dependencies: D.1.7
Description: Same as D.1.7 but for a set of root files. The output is txt file that is
use to feed Double_t e39times[50000] array in TbTaTool_files.h.
Table D.8: Function that matches a set of two points for various root files.

int i_index =0 beginning file index Loop for creating a root file int f_index =0
end file index
int fenergy = -404 energy
int fpolarity=-404 polarity
int which_tbegin=0
D.1.9 Generate a matched times analysis
Command name: GenerateMatchedTimesAnalysis()

Dependencies:D.1.7 and D.1.8
Description: Generate a analysis root file using D.1.7 and D.1.8 in order to study
if the set of points given as inputs to be matched have a average delta time
acceptable for the study.
Table D.9: Function that generate the plots that shows the interval between all the
matched points.
Input Description output

int which_tbegin Same condition for this input as D.1.6 Root file with the analysis of time
for matched points
D.1.10 Plotting one variable against all energies of the data

set
Command name: PlotOneVariableOverAllEnergies()

Description:
Generate histograms for the variable variable1 for all category1 = energies
and category2 = polarities.
Table D.10: Function that plots the histograms of one variable.

int variable_a_analizar =0 Since there are three types of results, this parameter Plots of the variable for all
indicates what tree is opened in TbTaTool::Init() the energies and polarities
char* variable1="delta_time" variable to plot
char* restriction_var ="delta_time = 0” ! restrictions to be applied to the variable
char* char* label_hist_result1 = "Title" title of the histogram
int nbins = 100 numbers of bins for the histograms
float xmin=0 bottom limit for the histogram
float xmax=200 upper limit for the histogram
D.1.11 Plotting stacked histograms for all energies
Command name: PlotStackAllEnergiesPionsTimeProfile()

Description:
Generate stacked histograms for the variable1 = Time for all the values of a
category == spills.
Table D.11: Plotting stacked histograms for all energies

int r =1 Since there are three types of results, this parameter Plots of the variable for all the energies
indicates tree is opened in TbTaTool::Init() and polarities stacked (r=1)
D.1.12 Plotting one variable for polarities

Command name: PlotOneVariableOnlyPolarity()
Description:
Generate all the plots for energy and polarity for a variable (e.g. duration_spill).
Table D.12: Plotting one variable for polarities

int variable_a_analizar Since there are three types of results, this parameter Plots for both polarities
indicates tree is opened in TbTaTool::Init()
char* variable1="my_cycle" variable to plot
char* restriction_var ="my_cycle = 0” ! restrictions to be applied to the variable
float xmin=60.2 bottom limit for the histogram
float xmax=60.9 upper limit for the histogram
D.1.13 Generating histograms for one variable for all energies

and stacked them in one plot
Command name: PlotOneVariableAllEnergiesStacked()

Description:
Generate histograms for the variable variable1 in all the energies and polarities
and then stack them into one plot.
Table D.13: Generating histograms for one variable for all energies and stacked them in
one plot

int variable_a_analizar =2 Since there are three types of results, this parameter Plots of the variable
indicates tree is opened in TbTaTool::Init() for all the energies
char* variable1="delta_time" variable to plot and polarities stacked
char* restriction_var ="delta_time = 0” ! restrictions to be applied to the variable in one plot
float xmin=60.2 bottom limit for the histogram
float xmax=60.9 upper limit for the histogram
D.1.14 GetValuesHistograms
Command name: GetValuesHistograms()

Description:
Gets the value of the events according with the conditions that generate the his-
tograms.
Table D.14: Match of to points.

int variable_a_analizar =0 Since there are three types of results, this parameter Plots of the variable
indicates tree is opened in TbTaTool::Init() for all the energies
char* variable1="delta_time" variable to plot and polarities
char* restriction_var ="delta_time = 0” ! restrictions to be applied to the variable
D.1.15 Production of all the analysis
Command name: GenerateDataPions()

Description:
Create, cut and plot all the variables that are presented in this thesis. This is
the analogous main() function in a usual C++ program, which contains all the
definitions and subrutines.
Appendix E
Developed Code
In this appendix I am showing the most important parts of code of the tool and
some auxiliary codes. For access to the full code, please go to https://github.
com/gsalazarq/TbTaTool/tree/master/TbTaTool
E.1 TbTaTool Time Profile (for only one subrun)
const char * name_file [1000] = { " / path / to / subrun1234 . root " }
// Value of $39 that match the begining of the subrun 1234

const Double_t * t_begin [1000] = { 1429844171}
void TimeToolForRun :: Loop ()

{
TString label = " p r o f i l e _ s p i l l s _ p i o n s _ r u n _ 2 . root " ;
char * name_results = label ;
// Creation of the ROOT File

TFile f_spill ( name_results , " RECREATE " );
TTree * tree_spill = new TTree ( " tree_spill " ," Tree Spill " );
Double_t Time_spill_1b , Time_spill_2b , Time_spill_3b ,

Time_spill_4b , Time_spill_5b , Time_spill_6b ;
139
APPENDIX E. DEVELOPED CODE 140
Int_t energyb , polarityb , Spill_numberb , file ;
TBranch * b_Time_spill_1b = tree_spill - > Branch ( " Time_spill_1b " ,

& Time_spill_1b , " Time_spill_1b / D " );
TBranch * b_file = tree_spill - > Branch ( " file " , & file , " file / I " );
TBranch * b_Spill_numberb = tree_spill - > Branch ( " Spill_numberb " ,
& Spill_numberb , " Spill_numberb / I " );
TBranch * b_energyb = tree_spill - > Branch ( " energyb " , & energyb , " energyb / I " );
TBranch * b_polarityb = tree_spill - > Branch ( " polarityb " , & polarityb , " polarityb / I " );
1
// Definition of auxiliary variables
Double_t duration_spill_1 , duration_spill_2 , duration_spill_3 ,

duration_spill_4 , duration_spill_5 , duration_spill_6 ;
Double_t t_o_spill_absolute , code_spill_1 , code_spill_2 ,
code_spill_3 , code_spill_4 , code_spill_5 , code_spill_6 ;
Double_t exists_spill_1 , exists_spill_2 , exists_spill_3 ,
exists_spill_4 , exists_spill_5 , exists_spill_6 ;
Double_t signal1_t_begin_spill_1 , signal1_t_begin_spill_2 ,
signal1_t_begin_spill_3 , signal1_t_begin_spill_4 , signal1_t_begin_spill_5 ,
signal1_t_begin_spill_6 ;
Double_t Time_spill , Time_spill_1 , Time_spill_2 , Time_spill_3 ,
Time_spill_4 , Time_spill_5 , Time_spill_6 ;
Double_t Time_begin_spill , Time_begin_spill_1 , Time_begin_spill_2 ,
Time_begin_spill_3 , Time_begin_spill_4 , Time_begin_spill_5 , T ime _b eg in _s pi ll _6 ;
Float_t i n t e r v a l _ b e t w e e n _ s p i l l s = 60.53333333;
TFile * f [10000];
TTree * tree [10000];
if ( fChain == 0) return ; Long64_t nentries = fChain - > GetEntries ();

s i g n a l 1 _ t _ b e g i n _ s p i l l _ 1 = 0; s i g n a l 1 _ t _ b e g i n _ s p i l l _ 2 = 0;
Ti me _b eg in _s pi ll _1 = -404; T ime _b eg in _s pi ll _2 = -404;

Long64_t ientry = LoadTree ( jentry ); // if ( ientry < 0) break ;
if ( In_spill > 0.5 ){ if ( Spill_number == 1 && exists_spill_1 != 0

){
Ti me _b eg in _s pi ll _1 = ( Double_t ) Time - t_begin [ i ];
s i g n a l 1 _ t _ b e g i n _ s p i l l _ 1 = 1;
break ; } }

){
break ; } }
}


){
break ; } }
}

// if ( ientry < 0) break ;
){
break ; } }
}

){
break ; } }
}


){
break ; } }
}
cout << " . " ; t _o _s pi ll _a bs ol ut e = 0;
code_spill_1 = - 404; code_spill_2 = - 404;

// = choose the absolute time considering the actual first spill that exist
if ( Tim e_ be gi n_ sp il l_ 1 != -404 ){ t_ o_s pi ll _a bs ol ut e = Ti me _be gi n_ sp il l_ 1 ;
code_spill_1 =100; }
else if ( T im e_ be gi n_ sp il l_ 2 != -404 ){ t_o _s pi ll _a bs ol ut e = Ti me_ be gi n_ sp il l_ 2 ;
// = Loop for cutting the Spills

if ( fChain == 0) return ; Long64_t nentries = fChain - > GetEntries ();

// Conditions for Spill has actual values

Time_spill_1b = -404; Time_spill_2b = -404; Time_spill_3b = -404;

if ( In_spill > 0.5 )

{
if ( Spill_number == 1 && exists_spill_1 != 0 )
{
Time_spill_1 = ( Double_t ) Time - t_begin [ i ];
Time_spill_1b = ( Double_t ) Time - t _o_ sp il l_ ab so lu te - t_begin [ i ] ;
Spill_numberb = 1;
}
else if ( Spill_number == 2 && exists_spill_2 != 0 )

{
if ( code_spill_1 != -404 ){
Time_spill_2b = ( Double_t ) Time - i n t e r v a l _ b e t w e e n _ s p i l l s *1
- t_begin [ i ] - t_o _s pi ll _a bs ol ut e ;
Spill_numberb = 2;
}
else {
Spill_numberb = 2;
}
}
{
if ( code_spill_1 != -404){
Spill_numberb = 3;
}
else if ( code_spill_2 != -404){
- t_begin [ i ] - t_ o_s pi ll _a bs ol ut e ;
Spill_numberb = 3;
}
else {
Spill_numberb = 3;
}
}
{
Spill_numberb = 4;
}
Spill_numberb = 4;
}
Spill_numberb = 4;
}
else {
Spill_numberb = 4;
}
}
{
Spill_numberb = 5;
}
Spill_numberb = 5;
}
Spill_numberb = 5;
}
Spill_numberb = 5;
}
else {
Spill_numberb = 5;
}
}
{
Spill_numberb = 6;
}
- t_begin [ i ] - t _o _s pi ll _a bs ol ut e ;
Spill_numberb = 6;
}
Spill_numberb = 6;
}
Spill_numberb = 6;
}
Spill_numberb = 6;
}
else {
Spill_numberb = 6;
}
} // end of else if ( Spill_number == 6 && exists_spill_6 != 0 )
} // end of : if ( In_spill > 0.5 )
} // end of : for ( Long64_t jentry =0; jentry < nentries ; jentry ++)
cout << " . " << endl ;
}
E.2 TbTaTool Spill Frequency and Spill Duration (for

only one subrun)
// ## 1
duration_spill = duration_spill_1 ;
category = 1;
mi_cycle = 0;
spill_global - > Fill ();
// ## 2
if ( s i g n a l 1 _ t _ b e g i n _ s p i l l _ 1 == 1 && s i g n a l 1 _ t _ b e g i n _ s p i l l _ 2 == 1){
mi_cycle = t_o_spill_2 - t_o_spill_1 ;
category = 1;
// myfile << mi_cycle << endl ;
// myfile4 << duration_spill << endl ;

}
else {
mi_cycle = 0;
category = 1;
}
E.2.1 Cut a variable of time into different parts according to a

criteria
for ( Long64_t jentry =0; jentry <1; jentry ++) {

Long64_t ientry = LoadTree ( jentry ); if ( ientry < 0) break ;
t_begin [ i ] = ( Double_t ) Time ;
// myfile2 << " t_begin : " < < t_begin [ i ] << endl ; }
// LOOP 1 : Begin Time for the Spill_number == 1 //
if ( In_spill > 0.5 ){ if ( Spill_number == 1 && exists_spill_1 != 0 ){

// jentry_array = ( Double_t ) jentry ;
Time_begin_spill_1 = ( Double_t ) Time - t_begin [ i ];

t_o_spill_1 = ( Double_t ) Time ;
signal1_t_begin_spill_1 = 1;
break ; } } }
E.2.2 Calculation of different variables regarding time


// Conditions for Spill has actual values

mi_cycle = 0;
if ( In_spill > 0.5 ) {

if ( Spill_number == 1 ) {
duration_spill_1 = Time_spill_1 - Time_begin_spill_1 ;
Time_spill_1b = ( Double_t ) Time - t_o_spill_1 ;
Spill_numberb = 1;
}
else if ( Spill_number == 2 )
{
duration_spill_2 = Time_spill_2 - Time_begin_spill_2 ;
Time_spill_2b = ( Double_t ) Time - t_o_spill_2 ;
Spill_numberb = 2;
}
E.3 Auxiliary Tools
E.3.1 Tool for match the $39 signal and the corresponding
root file
if ( tree_e39 == 0) return ;
Long64_t nentries_e39 = tree_e39 - > GetEntries ();

Long64_t nbytes_e39 = 0 , nb_e39 = 0;
for ( Long64_t jentry_e39 =0; jentry_e39 < nentries_e39 ; jentry_e39 ++) {

Long64_t ientry_e39 = LoadTree ( jentry_e39 );
// if ( ientry_e39 < 0) break
nb = tree_e39 - > GetEntry ( jentry_e39 ); nbytes_e39 += nb_e39 ;

// if ( Cut ( ientry ) < 0) continue ;
delta_time = t_begin [ i ] - unixtime_e39 ;
if ( delta_time > 0 ){ }
else {
// myfile << delta_time_old << endl ;
tree_diff - > Fill ();
break ;
}
delta_time_old = delta_time ; time_root_file = t_begin [ i ];

last_jentry_e39 = jentry_e39 ;
}
file = i ;
// tree_diff - > Fill ();
cout << i << " . " << endl ;
E.3.2 Conversion between Unixtime into readable human time
Unix machines use the unixtime as the way to tag the events. This tools is useful
since e39 data come into date-like format, while time in the data is in unixtime.
The time of the ROOT files are 5 hours before the Central Time, which is the time
at Fermilab. This correction was introduce in the data.
This code convert unixtime into human-readable date:
# !/ usr / bin / python
import datetime
f = open ( " MINERVA_E39 . txt " ," a + " )
f_converted = open ( " unixtime_MINERVA_E39 . txt " ," a + " )
print " File opened "
array_months = [ " MAR " ," APR " ]
# for reading lines from a file
i = 0
for line in f :
day = line [0:2]
month = line [3:6]
year = line [7:11]
mi_cicle_line_1 = line [25:27]
mi_cicle_line_2 = line [28:]
mi_cicle_line = mi_cicle_line_1 + " . " + mi_cicle_line_2
d = int ( day )
y = int ( year )
if month == array_months [0]:

month_n = 3
else :
month_n = 4
m = int ( month_n )
time_line = datetime . datetime (y ,m ,d ,0 , 0 , 0)
unix_reference = datetime . datetime (1970 ,1 ,1 , 0 , 0 , 0)
delta_time = time_line - unix_reference
days_delta = delta_time . days
hours = line [12:14]

h = float ( hours )
minutes = line [15:17]
m = float ( minutes )
seconds = line [18:24]
s = float ( seconds )
# calculation of the seconds

# unixtime_line = s + m *60 + h *3600 + days_delta *86400
unixtime_line = s + m *60 + h *3600 + days_delta *86400 + 5*3600
f_converted . write ( " % s \ n " % str ( unixtime_line ) )
i = i +1
print i
# f_converted . write ("% s " % mi_cicle_line )
The output of this code show the unixtime and the date in UCT
UNIX Time Date

1426939223 23 - APR -2015 _20 :00:09.669
1426956777 23 - APR -2015 _20 :01:10.234
1426957317 23 - APR -2015 _20 :02:10.799
continues ....
E.4 Election of the reference point for time profile
// == Computing Delta Interval e39 ==
if ( tree_e39 == 0) return ;
Long64_t nentries_e39 = tree_e39 - > GetEntries ();
Long64_t nbytes_e39 = 0 , nb_e39 = 0;
for ( Long64_t jentry_e39 =0; jentry_e39 < nentries_e39 ; jentry_e39 ++) {
Long64_t ientry_e39 = LoadTree ( jentry_e39 );

nb = tree_e39 - > GetEntry ( jentry_e39 ); nbytes_e39 += nb_e39 ;
delta_time = t_begin [ i ] - unixtime_e39 ;
if ( t_begin [ i ] != -404){
if ( delta_time > 0 ){ }
else {
tree_diff - > Fill ();
break ;}}
else {
delta_time = -404;
tree_diff - > Fill ();}
delta_time_e39 = delta_time ;
time_root_file = t_begin [ i ];
time_unixtimee39 = unixtime_e39 ;
file = i ;}
E.5 Creation of ROOT Files from TXT files
// reading a text file

# include < iostream >
# include < fstream >
# include < string >
using namespace std ;
void ReadAsciiCreateRootFile (){
TFile * f = new TFile ( " temp . root " ," RECREATE " );
TTree * tree_mi_cycle = new TTree ( " tree_mi_cycle " ," Tree MI Cycle " );
Double_t time_mi_cycle = 0 , mi_cycle = 0 ;
TBranch * b_time_mi_cycle = tree_mi_cycle - > Branch ( " time_mi_cycle " ,
& time_mi_cycle , " time_mi_cycle / D " );
TBranch * b_mi_cycle = tree_mi_cycle - > Branch ( " mi_cycle " , & mi_cycle ,
" mi_cycle / D " );
string line ;
ifstream myfile ( " data2 " );
if ( myfile . is_open ()){
while ( getline ( myfile , line ))
{
cout << line << " \ n " ;
in >> time_mi_cycle >> mi_cycle ; // format of the data
tree_mi_cycle - > Fill ();
}
myfile . close ()
}
else cout << " Unable to open the file " << endl ;
for ( Int_t i ; i < nlines ; i ++){
if (! in . good ())
// if ( nlines <5) printf (" time_mi_cycle =%8 f ,
mi_cycle =%8 f " , time_mi_cycle , mi_cycle );
}
printf ( " found % d points \ n " , nlines );
in . close ();
f - > Write ();
}
Bibliography
[1] E. Kh. Akhmedov. The neutrino magnetic moment and time variations of the
solar neutrino flux. page 22, may 1997.
[2] L Aliaga and MINERvA Collaboration. Design, Calibration, and Performance

of the MINERvA Detector. Nucl. Instrum. Methods A, may 2013.
[3] L Aliaga and MINERvA Collaboration. MINERvA neutrino detector response

measured with test beam data. Nuclear Instruments and Methods in Physics
Research Section A: Accelerators, Spectrometers, Detectors and Associ-
ated Equipment, 789:1–44, jan 2015.
[4] Ilka Antcheva, Maarten Ballintijn, Bertrand Bellenot, Marek Biskup, Rene
Brun, Nenad Buncic, Philippe Canal, Diego Casadei, Olivier Couet, Valery
Fine, Leandro Franco, Gerardo Ganis, Andrei Gheata, David Gonzalez Ma-
line, Masaharu Goto, Jan Iwaszkiewicz, Anna Kreshuk, Diego Marcos Se-
gura, Richard Maunder, Lorenzo Moneta, Axel Naumann, Eddy Offermann,
Valeriy Onuchin, Suzanne Panacek, Fons Rademakers, Paul Russo, and
Matevz Tadel. ROOT - A C++ Framework for Petabyte Data Storage, Statis-
tical Analysis and Visualization. (September 2015):1–33, aug 2015.
[5] Michael Backfish, Leo Bellantoni, Anne Norrick, and Geoff Savage. Time Of
Flight for MINERvA Testbeam II. Technical report, 2014.
[6] N. J. Baker, P. L. Connolly, S. A. Kahn, M. J. Murtagh, R. B. Palmer,

N. P. Samios, and M. Tanaka. otal cross sections for $ {} n$ and $ {}p$
charged-current interactions in the 7-foot bubble chamber. Physical Review
D, 25(3):617–623, feb 1982.
156
BIBLIOGRAPHY 157
[7] D Casper. The nuance neutrino physics simulation, and the future. Nuclear
Physics B - Proceedings Supplements, 112(1-3):161–170, nov 2002.
[8] J. Chadwick. Possible Existence of a Neutron. Nature, 129(3254):402–402,

mar 1932.
[9] Daniel Cherdack and Elizabeth Worcester. Summary of long-baseline sys-

tematics session at CETUP*2014. page 030001, jan 2015.
[10] Bruce T. Cleveland, Timothy Daily, Raymond Davis, Jr., James R. Distel,
Kenneth Lande, C. K. Lee, Paul S. Wildenhain, and Jack Ullman. Mea-
surement of the Solar Electron Neutrino Flux with the Homestake Chlorine
Detector. The Astrophysical Journal, 496(1):505–526, mar 1998.
[11] Super-Kamiokande Collaboration. A Measurement of Atmospheric Neu-

trino Oscillation Parameters by Super-Kamiokande I. Physical Review D,
71(11):112005, jan 2005.
[12] Accelerator Division. Concepts Rookie Book. 2013.
[13] J. a. Formaggio and G. P. Zeller. From eV to EeV: Neutrino Cross Sections

Across Energy Scales. Reviews of Modern Physics, 84(3):1307–1341, may
2013.
[14] Y. Fukuda, T. Hayakawa, and SuperKamiokande Collaboration. Evidence for

Oscillation of Atmospheric Neutrinos. Physical Review Letters, 81(8):1562–
1567, aug 1998.
[15] Thomas K Gaisser. Atmospheric Neutrino Fluxes. Physica Scripta,

T121:51–56, feb 2005.
[16] David J. Griffiths. Introduction to Elementary Particles, 2008.
[17] Deborah a. Harris. The State of the Art of Neutrino Cross Section Measure-
ments. Prospects in Neutrino Physics Conference, page 11, jun 2015.
[18] K. S. Hirata, K. Inoue, T. Ishida, T. Kajita, K. Kihara, M. Nakahata, K. Naka-

mura, S. Ohara, A. Sakai, N. Sato, Y. Suzuki, Y. Totsuka, Y. Yaginuma,
BIBLIOGRAPHY 158
M. Mori, Y. Oyama, A. Suzuki, K. Takahashi, M. Yamada, M. Koshiba,

K. Nishijima, T. Kajimura, T. Suda, T. Tajima, K. Miyano, H. Miyata, H. Takei,
Y. Fukuda, E. Kodera, Y. Nagashima, M. Takita, H. Yokoyama, K. Kaneyuki,
Y. Takeuchi, T. Tanimori, E. W. Beier, E. D. Frank, W. Frati, S. B. Kim, A. K.
Mann, F. M. Newcomer, R. Van Berg, and W. Zhang. Observation of a small
atmospheric v??/ve ratio in Kamiokande. Physics Letters B, 280(1-2):146–
152, apr 1992.
[19] Edward Kearns. Experimental measurements of atmospheric neutrinos. Nu-

clear Physics B - Proceedings Supplements, 70(1-3):315–323, jan 1999.
[20] T. Leitner. Neutrino Interactions with Nucleons and Nuclei. Diplomat, Justus-
Liebig-Universität Gießen, 2005.
[21] Andrew John Lowe. Neutrino Physics & The Solar Neutrino Problem.
page 40, jul 2009.
[22] Marco Martini. Recent developments in neutrino-nucleus scattering theory.

In Morfı́n J. da Motta H., editor, NuFact15 : XVII International Workshop on
Neutrino Factories and Future Neutrino Facilities, Rio de Janeiro, 2015.
[23] R. N. Mohapatra, S. Antusch, K. S. Babu, G. Barenboim, M-C Chen,

A de Gouvêa, P. de Holanda, B. Dutta, Y. Grossman, A. Joshipura,
B. Kayser, J. Kersten, Y. Y. Keum, S. F. King, P. Langacker, M. Lindner,
W. Loinaz, I. Masina, I. Mocioiu, S. Mohanty, H. Murayama, S. Pascoli,
S. T. Petcov, A. Pilaftsis, P. Ramond, M. Ratz, W. Rodejohann, R. Shrock,
T. Takeuchi, T. Underwood, and L. Wolfenstein. Theory of neutrinos: a white
paper. Reports on Progress in Physics, 70(11):1757–1867, nov 2007.
[24] Mihoko M. Nojiri. Beyond the Standard Model. jun 2014.
[25] Anne Norrick. The Minerva Detector. In MINERVA 101, 2015.
[26] K.A. Olive. Review of Particle Physics. Chinese Physics C, 38(9):090001,

aug 2014.
[27] Cheryl Patrick. Calculation cross-sections. 2015.

BIBLIOGRAPHY 159
[28] Carlos Pena-Garay and Aldo Serenelli. Solar neutrinos and the solar com-
position problem. page 5, nov 2008.
[29] G N Perdue, L Bagby, and MINERvA Collaboration. The data acquisition

system and infrastructure. Nuclear Instruments and Methods in Physics Re-
search Section A: Accelerators, Spectrometers, Detectors and Associated
Equipment, 694(0):179–192, 2012.
[30] M. L. Perl, G. S. Abrams, A. M. Boyarski, M. Breidenbach, D. D. Briggs, F. Bu-

los, W. Chinowsky, J. T. Dakin, G. J. Feldman, C. E. Friedberg, D. Fryberger,
G. Goldhaber, G. Hanson, F. B. Heile, B. Jean-Marie, J. A. Kadyk, R. R.
Larsen, A. M. Litke, D. L??ke, B. A. Lulu, V. L??th, D. Lyon, C. C. Morehouse,
J. M. Paterson, F. M. Pierre, T. P. Pun, P. A. Rapidis, B. Richter, B. Sadoulet,
R. F. Schwitters, W. Tanenbaum, G. H. Trilling, F. Vannucci, J. S. Whitaker,
F. C. Winkelmann, and J. E. Wiss. Evidence for anomalous lepton produc-
tion in e+-e- annihilation. Physical Review Letters, 35(22):1489–1492, dec
1975.
[31] Frederick Reines and Clyde L. Cowan. The Neutrino. Nature, 178(Septem-
ber):3, 1956.
[32] U Tokyo and Particle Data Group Group. Neutrino mass, mixing, and oscil-
lations, 2014.
[33] Geralyn P. Zeller, Jorge G. Morfin, and F. (Flavio) Cavanna. Neutrino-nucleus

interactions in the few-gev region : NuInt07, the 5th International Workshop
on Neutrino-Nucleus Interactions in the Few-GeV Region, Batavia Illinois,
30 May-3 June 2007, volume 967. American Institute of Physics, 2007.

Salazar QG

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Salazar QG

Uploaded by

Copyright:

Available Formats

UNIVERSIDAD NACIONAL DE INGENIERÍA

PARA OBTENER EL GRADO ACADÉMICO DE MAESTRO EN

GERALD FERNANDO SALAZAR QUIROZ

Dr. CARLOS JAVIER SOLANO SALINAS

Dedicado a Fisher, Arnold, Ana, Nataly, Abel y Carmen.

Do not go gentle into that good night,

Though wise men at their end know dark is right,

Dylan Thomas, 1914 - 1953

3 MINERvA a cross-section neutrino experiment 29

4 Time structure of the FBTF Beam 54

5 Tools for Data Analysis 71

A Auxiliary plots 100

B Participation in the MINERνA experiment during 2015 109

C Radiofrequency Cavities theory 119

D Documentation of Code 132

D.1.6 Generation of txt file with timestamps of beginning of the

E Developed Code 139

2.1 Total cross-sections for Charge Current Quasi-elastic Scattering

2.4 All νµ quasi-elastic scattering cross-sections, νµ n → µ− p measure-

3.1 Front view of the MINERvA detector at NuMI Hall at Fermilab. . . . 30

4.1 Fermilab Accelerator Complex. . . . . . . . . . . . . . . . . . . . . 58

5.1 A standardd histogram with box plots generate in ROOT. . . . . . . 73

6.1 Time profile for Run2 ±4, ±6 and ±8 GeV (pions). . . . . . . . . . 90

A.1 Spill frequency for 4GeV Pions . . . . . . . . . . . . . . . . . . . . 100

A.15 Spill duration for electrons (Run3). . . . . . . . . . . . . . . . . . . 107

C.1 Examples of wave guides. . . . . . . . . . . . . . . . . . . . . . . . 123

1.1 Range of interaction of the four fundamental forces. . . . . . . . . 6

2.1 Experiments focus on cross-sections and oscillation of neutrinos

4.1 Time structure of the Beam according to Accelerator Division. . . . 67

D.1 Calculation of the number of events in one subrun. . . . . . . . . . . . 132

Tı́tulo: Charge pion production in interactions on hydrocarbon at <E>!=

(b) Physical Review Letters Vol 116 (2016) 081802

Finally in Appendix B, a review of other activities during the intership at Fermi-

One of the greatest achievements in science, is the formulation of the Standard

”For the discovery of neutrino oscillation which shows that neutrinos

Other neutrino Nobel Prizes have been awarded in 1988 to L. M. Lederman,

1.1 Weak Interactions

Figure 1.2: Fundamental vertex for weak interactions.

A more complicated reactions can be produce if the primitive diagram is com-

π − → µ− ν¯µ , with τ = 2.6 × 10−8 sec,

Interaction Range Lifetime Cross Section Coupling

1.2 Neutrino Physics

Neutrino was proposed by Wolfgang Pauli in 1930 to hold up the energy-momentum

In 1932, a more massive neutral particle was discovered by Chadwick([8]),

Figure 1.3: Schematic diagram of neutrino detector as appeared in the original

Γ(π + → µ+ νL ) 6= Γ(π + → µ+ νR ) P violation,

but the CP variance is conserved. Future experiments have as main objective

Γ(π + → µ+ νL ) = Γ(π − → µ− ν¯R ) CP invariance. (1.7)

1.2.1 Solar and Atmospheric Neutrinos

At low energies (≤ 1GeV ), the ratio between the flux of muon-neutrinos to

If the flux of neutrinos coming from the atmosphere is expected to be isotropic,

1.2.2 Neutrino oscillations

|νe i = |ν1 icosθ + |ν2 isinθ

|νt i = −|ν1 ie−iE1 t sinθ + |ν2 ie−iE2 t cosθ, (1.10)

P (νµ → νe ; t) = |hνe |ν(t)i|2

Table 1.2: Sensitivity of different oscillation experiments.[32, p. 11]

Source Type of ν E[MeV] L[km] min(∆m2 )[eV 2 ]

Reactor ν¯e ∼1 1 ∼ 10−3

MINERνA3 is a cross-sections precision studies experiment of (anti)neutrino-

2.1 Neutrino interaction with matter

Number of reactions of a given type per unit time

σ is a number that has dimensions of area and is usually expressed in cm2 or

Figure 2.1: Total cross-sections for Charge Current Quasi-elastic Scattering on